A Rosetta Stone For Quantum Mech. W. An Intro To Quantum Computation (JNL Article) - S. Lomonaco (2000) WW PDF

arXiv:quant-ph/0007045 v1 17 Jul 2000
A ROSETTA STONE FOR QUANTUM MECHANICS WITH

AN INTRODUCTION TO QUANTUM COMPUTATION
VERSION 1.5
SAMUEL J. LOMONACO, JR.
Abstract. The purpose of these lecture notes is to provide readers,

who have some mathematical background but little or no exposure to
quantum mechanics and quantum computation, with enough material
to begin reading the research literature in quantum computation and
quantum information theory. This paper is a written version of the
first of eight one hour lectures given in the American Mathematical Society (AMS) Short Course on Quantum Computation held in conjunction
with the Annual Meeting of the AMS in Washington, DC, USA in January 2000, and will be published in the AMS PSAPM volume entitled
Quantum Computation..
Part 1 of the paper is a preamble introducing the reader to the concept of the qubit,
Part 2 gives an introduction to quantum mechanics covering such topics as Dirac notation, quantum measurement, Heisenberg uncertainty,
Schr
odingers equation, density operators, partial trace, multipartite
quantum systems, the Heisenberg versus the Schr
odinger picture, quantum entanglement, EPR paradox, quantum entropy.
Part 3 gives a brief introduction to quantum computation, covering such topics as elementary quantum computing devices, wiring diagrams, the no-cloning theorem, quantum teleportation, Shors algorithm, Grovers algorithm.
Many examples are given to illustrate underlying principles. A table of contents as well as an index are provided for readers who wish
to pick and choose. Since this paper is intended for a diverse audience, it is written in an informal style at varying levels of difficulty and
sophistication, from the very elementary to the more advanced.
Date: June 20, 2000.

2000 Mathematics Subject Classification. Primary: 81-01, 81P68.
Key words and phrases. Quantum mechanics, quantum computation, quantum algorithms, entanglement, quantum information.
This work was partially supported by ARO Grant #P-38804-PH-QC and the L-O-O-P
Fund. The author gratefully acknowledges the hospitality of the University of Cambridge
Isaac Newton Institute for Mathematical Sciences, Cambridge, England, where some of
this work was completed.
I would also like to thank the other AMS Short Course
lecturers, Howard Brandt, Dan Gottesman, Lou Kauffman, Alexei Kitaev, Peter Shor,
Umesh Vazirani and the many Short Course participants for their support. (Copyright
2000.)
1
Contents
Part 1.
Preamble
1. Introduction
2. The classical world

2.1. Introducing the Shannon bit.
2.2. Polarized light: Part I. The classical perspective
5
5
5
3. The quantum world

3.1. Introducing the qubit But what is a qubit?
3.2. Where do qubits live? But what is a qubit?
3.3. A qubit is ...
6
6
7
7
Part 2.
An Introduction to Quantum Mechanics
4. The beginnings of quantum mechanics

4.1. A Rosetta stone for Dirac notation: Part I. Bras, kets, and
bra-(c)-kets
4.2. Quantum mechanics: Part I. The state of a quantum system
4.3. A Rosetta stone for Dirac notation: Part II. Operators
4.4. Quantum mechanics: Part II. Observables
4.5. Quantum mechanics: Part III. Quantum measurement General
principles
4.6. Polarized light: Part III.
Three examples of quantum
measurement
4.7. A Rosetta stone for Dirac notation: Part III. Expected values
4.8. Quantum Mechanics: Part IV. The Heisenberg uncertainty
principle
4.9. Quantum mechanics: Part V. Dynamics of closed quantum
systems: Unitary transformations, the Hamiltonian, and
Schr
odingers equation
4.10. The mathematical perspective
5. The Density Operator
5.1. Introducing the density operator
5.2. Properties of density operators
5.3. Quantum measurement in terms of density operators
5.4. Some examples of density operators
5.5. The partial trace of a linear operator
5.6. Multipartite quantum systems
5.7. Quantum dynamics in density operator formalism
5.8. The mathematical perspective
8
8
8
10
12
13
15
16
17
18
19
19
20
20
21
22
23
24
26
27
27
A ROSSETTA STONE FOR QUANTUM MECHANICS
6. The Heisenberg model of quantum mechanics
28
7. Quantum entanglement
31
7.1. The juxtaposition of two quantum systems
31
7.2. An example: An n-qubit register Q consisting of the juxtaposition
of n qubits.
32
7.3. An example of the dynamic behavior of a 2-qubit register
33
7.4. Definition of quantum entanglement
35
7.5. Einstein, Podolsky, Rosens (EPRs) grand challenge to quantum
mechanics.
36
7.6. Why did Einstein, Podolsky, Rosen (EPR) object?
37
7.7. Quantum entanglement: The Lie group perspective
39
8. Entropy and quantum mechanics
8.1. Classical entropy, i.e., Shannon Entropy
8.2. Quantum entropy, i.e., Von Neumann entropy
8.3. How is quantum entropy related to classical entropy?
8.4. When a part is greater than the whole Ignorance = uncertainty
43
43
44
46
46
9. There is much more to quantum mechanics
48
Part of a Rosetta Stone for Quantum

Computation
Part 3.
48
10. The Beginnings of Quantum Computation - Elementary

Quantum Computing Devices
10.1. Embedding classical (memoryless) computation in quantum
mechanics
10.2. Classical reversible computation without memory
10.3. Embedding classical irreversible computation within classical
reversible computation
10.4. The unitary representation of reversible computing devices
10.5. Some other simple quantum computing devices
10.6. Quantum computing devices that are not embeddings
10.7. The implicit frame of a wiring diagram
51
51
53
54
54
11. The No-Cloning Theorem
55
12. Quantum teleportation
57
13. Shors algorithm

13.1. Preamble to Shors algorithm
13.2. Number theoretic preliminaries
13.3. Overview of Shors algorithm
13.4. Preparations for the quantum part of Shors algorithm
13.5. The quantum part of Shors algorithm
60
61
62
62
64
65
48
49
49
13.6.
13.7.
13.8.
13.9.
13.10.
Peter Shors stochastic source S

A momentary digression: Continued fractions
Preparation for the final part of Shors algorithm
The final part of Shors algorithm
An example of Shors algorithm
67
69
70
75
76
14. Grovers Algorithm

14.1. Problem definition
14.2. The quantum mechanical perspective
14.3. Properties of the inversion I|i
14.4. The method in Luvs madness
14.5. Summary of Grovers algorithm
14.6. An example of Grovers algorithm
80
80
81
82
83
87
88
15. There is much more to quantum computation
90
References
Index
90
95
Part 1.
Preamble
1. Introduction
These lecture notes were written for the American Mathematical Society
(AMS) Short Course on Quantum Computation held 17-18 January 2000
in conjunction with the Annual Meeting of the AMS in Washington, DC in
January 2000. The notes are intended for readers with some mathematical
background but with little or no exposure to quantum mechanics. The
purpose of these notes is to provide such readers with enough material in
quantum mechanics and quantum computation to begin reading the vast
literature on quantum computation, quantum cryptography, and quantum
information theory.
The paper was written in an informal style. Whenever possible, each
new topic was begun with the introduction of the underlying motivating
intuitions, and then followed by an explanation of the accompanying mathematical finery. Hopefully, once having grasped the basic intuitions, the
reader will find that the remaining material easily follows.
Since this paper is intended for a diverse audience, it was written at
varying levels of difficulty and sophistication, from the very elementary to
the more advanced. A large number of examples have been included. An
index and table of contents are provided for those readers who prefer to
pick and choose. Hopefully, this paper will provide something of interest
for everyone.
Because of space limitations, these notes are, of necessity, far from a
complete overview of quantum mechanics. For example, only finite dimensional Hilbert spaces are considered, thereby avoiding the many pathologies
that always arise when dealing with infinite dimensional objects. Many
important experiments that are traditionally part of the standard fare in
quantum mechanics texts (such as for example, the Stern-Gerlach experiment, Youngs two slit experiment, the Aspect experiment) have not been
mentioned in this paper. We leave it to the reader to decide if these notes
have achieved their objective.
The final version of this paper together with all the other lecture notes
of the AMS Short Course on Quantum Computation will be published as a
book in the AMS PSAPM Series entitled Quantum Computation.
2. The classical world

2.1. Introducing the Shannon bit.
Since one of the objectives of this paper is to discuss quantum information,
we begin with a brief discussion of classical information.
The Shannon bit is so well known in our age of information that it needs
little, if any, introduction. As we all know, the Shannon bit is like a very
decisive individual. It is either 0 or 1, but by no means both at the same
time. The Shannon bit has become so much a part of our every day lives
that we take many of its properties for granted. For example, we take for
granted that Shannon bits can be copied.
2.2. Polarized light: Part I. The classical perspective.
Throughout this paper the quantum polarization states of light will be

used to provide concrete illustrations of underlying quantum mechanical
principles. So we also begin with a brief discussion of polarized light from
the classical perspective.
Light waves in the vacuum are transverse electromagnetic (EM) waves
with both electric and magnetic field vectors perpendicular to the direction
of propagation and also to each other. (See figure 1.)
Figure 1. A linearly polarized electromagnetic wave.

If the electric field vector is always parallel to a fixed line, then the EM
wave is said to be linearly polarized. If the electric field vector rotates
about the direction of propagation forming a right-(left-)handed screw, it is
said to be right (left) elliptically polarized. If the rotating electric field
vector inscribes a circle, the EM wave is said to be right-or left-circularly
polarized.
3. The quantum world
3.1. Introducing the qubit But what is a qubit?
Many of us may not be as familiar with the quantum bit of information,
called a qubit. Unlike its sibling rival, the Shannon bit, the qubit can be
both 0 and 1 at the same time. Moreover, unlike the Shannon bit, the
qubit can not be duplicated1 . As we shall see, qubits are like very slippery,
irascible individuals, exceedingly difficult to deal with.
One example of a qubit is a spin 21 particle which can be in a spin-up state
|1i which we label as 1, in a spin-down state |0i which we label as 0, or
in a superposition of these states, which we interpret as being both 0 and
1 at the same time. (The term superposition will be explained shortly.)
Another example of a qubit is the polarization state of a photon. A
photon can be in a vertically polarized state |li. We assign a label of 1
to this state. It can be in a horizontally polarized state |i. We assign a
label of 0 to this state. Or, it can be in a superposition of these states.
In this case, we interpret its state as representing both 0 and 1 at the same
time.
Anyone who has worn polarized sunglasses is familiar with the polarization states of light. Polarized sunglasses eliminate glare by letting through
only vertically polarized light, while filtering out the horizontally polarized
1
This is a result of the no-cloning theorem of Wootters and Zurek[83]. A proof of the
no-cloning theorem is given in Section 10.8 of this paper.
light. For that reason, they are often used to eliminate road glare, i.e.,
horizontally polarized light reflected from the road.
3.2. Where do qubits live? But what is a qubit?
But where do qubits live? They live in a Hilbert space H. By a Hilbert
space, we mean:
A Hilbert space H is a vector space over the complex numbers C with
a complex valued inner product
(, ) : H H C
which is complete with respect to the norm
p
kuk = (u, u)
induced by the inner product.
Remark 1. By a complex valued inner product, we mean a map

(, ) : H H C
from H H into the complex numbers C such that:

1) (u, u) = 0 if and only if u = 0
2) (u, v) = (v, u)
3) (u, v + w) = (u, v) + (u, w)
4) (u, v) = (u, v)
where denotes the complex conjugate.
Remark 2. Please note that (u, v) = (u, v).
3.3. A qubit is ...
A qubit is a quantum system Q whose

state lies in a two dimensional Hilbert space H.
2
Barenco et al in [1] define a qubit as a quantum system with a two dimensional Hilbert
space, capable of existing in a superposition of Boolean states, and also capable of being
entangled with the states of other qubits. Their more functional definition will take on
more meaning as the reader progresses through this paper.
Part 2.
An Introduction to Quantum Mechanics
4. The beginnings of quantum mechanics

4.1. A Rosetta stone for Dirac notation: Part I. Bras, kets, and
bra-(c)-kets .
The elements of a Hilbert space H will be called ket vectors, state
kets, or simply kets. They will be denoted as:
where label denotes some label.
| label i
Let H denote the Hilbert space of all Hilbert space morphisms of H into
the Hilbert space of all complex numbers C, i.e.,
H = HomC (H, C) .
The elements of H will be called bra vectors, state bras, or simply bras.
They will be denoted as:
h label |
where once again label denotes some label.
Also please note that the complex number

will simply be denoted by
h label1 | (| label2 i)
h label1 | label2 i
and will be called the bra-(c)-ket product of the bra h label1 | and the ket
| label2 i.
There is a monomorphism (which is an isomorphism if the underlying

Hilbert space is finite dimensional)
defined by
H H
| label i 7 ( | label i , )
The bra ( | label i , ) is denoted by h label |.
Hence,
h label1 | label2 i = (| label1 i , | label2 i)
Remark 3. Please note that ( | label i) = hlabel|.
The tensor product3 H K of two Hilbert spaces H and K is simply
the simplest Hilbert space such that
1) (h1 + h2 ) k = h1 k + h2 k, for all h1 , h2 H and for all k K,
and
2) h (k1 + k2) = h k1 + h k2 for all h H and for all k1 , k2 K.
3) (h k) (h) k = h (k) for all C, h H, k K.
Remark 4. Hence, k |labeli k =
(| label1 i , | label2 i) .
p
h label | label i and h label1 | label2 i =
It follows that, if { e1 , e2, . . . , em } and { f1 , f2 , . . . , fn } are respectively

bases of the Hilbert spaces H and K, then { ei fj | 1 i m, 1 j n }
is a basis of H K. Hence, the dimension of the Hilbert space H K is the
product of the dimensions of the Hilbert spaces H and K, i.e.,
Dim (H K) = Dim (H) Dim (K) .
Finally, if | label1 i and | label2 i are kets respectively in Hilbert spaces H1

and H2 , then their tensor product will be written in any one of the following
three ways:
| label1 i | label2 i
| label1 i | label2 i
| label1 , label2 i
3
Readers well versed in homological algebra will recognize this informal definition as a
slightly disguised version of the more rigorous universal definition of the tensor product.
For more details, please refer to [14], or any other standard reference on homological
algebra.
10
4.2. Quantum mechanics: Part I. The state of a quantum system.
The states of a quantum system Q are represented by state kets in a

Hilbert space H. Two kets |i and |i represent the same state of a quantum
system Q if they differ by a non-zero multiplicative constant. In other words,
|i and |i represent the same quantum state Q if there exists a non-zero
C such that
|i = |i
Hence, quantum states are simply elements of the manifold
H/ = CP n1
where n denotes the dimension of H, and CP n1 denotes complex projective (n 1)-space .

Convention: Since a quantum mechanical state is represented by a state
ket up to a multiplicative constant, we will, unless stated otherwise,
choose those kets |i which are of unit length, i.e., such that
h | i = 1 k |ik = 1
4.2.1. Polarized light: Part II. The quantum mechanical perspective.

As an illustration of the above concepts, we consider the polarization
states of a photon.
The polarization states of a photon are represented as state kets in a two
dimensional Hilbert space H. One orthonormal basis of H consists of the
kets
| i and |i
which represent respectively the quantum mechanical states of left- and
right-circularly polarized photons. Another orthonormal basis consists of
the kets
|li and |i
representing respectively vertically and horizontally linearly polarized photons. And yet another orthonormal basis consists of the kets
|%i and |&i
for linearly polarized photons at the angles = /4 and = /4 off the
vertical, respectively.
These orthonormal bases are related as follows:
|%i =
|%i = 2 (|li + |i)
|&i =
|li =
|i =
|i =
| i =
1
2
1
2
(|li |i)
1
2
(|%i |&i)
|&i =
|li =
(|%i + |&i)
1
2
(|li i |i)
1
2
(|li + i |i)
1+i
2
|i +
1i
2
| i
1i
2
|i +
1+i
2
| i
|i =
|i =
11
| i =
1
2
(|i + | i)
(|i | i)
1i
2
|%i +
1+i
2
|&i
1+i
2
|%i +
1i
2
|&i
The bracket products of the various polarization kets are given in the
table below:
|li |i |%i |&i
|i | i
1
1
1
1
hl|
1
0
2
2
2
2
1
1
i
i
h| 0
1
2
2
2
2
h%|
h&|
h|
h |
1
2
1
2
1
2
1
2
1
2
12
i
2
i2
1
0
0
1
1i
2
1+i
2
1+i
2
1i
2
1+i
2
1i
2
1i
2
1+i
2
1
0
0
1
In terms of the basis {|li , |i} and the dual basis {hl| , h|}, these kets
and bras can be written as matrices as indicated below:

1
1 0 ,
hl| =
|li =
h| =
h%| =
h&| =
h| =
h | =
0 1
1
2
|i =
1 1
1
2
1 1
1
2
1 i
1
2
1 i
|%i =
|&i =
|i =
| i =
1
2
0
1
1
1
1
1

1
1
2
i
1
2
1
2
In this basis, for example, the tensor product |%i is
1
i
12
|%i =
1
2
1
2
1
2
i2
1
1 i
=
2 1
i
and the projection operator | i h | is:

1 1 i
1
1
1
1 i =
| i h | =
i
1
i
2
2
2
4.3. A Rosetta stone for Dirac notation: Part II. Operators.
An (linear) operator or transformation O on a ket space H is a
Hilbert space morphism of H into H, i.e., is an element of
HomC (H, H)
The adjoint O of an operator O is that operator such that

O | label1 i , | label2 i = (| label1 i , O | label2 i)
for all kets | label1 i and | label2 i.
In like manner, an (linear) operator or transformation on a bra space H

is an element of
HomC (H , H )
Moreover, each operator O on H can be identified with an operator, also
denoted by O, on H defined by
h label1 | 7 h label1 | O
where h label1 | O is the bra defined by
(h label1 | O) (| label2i) = h label1 | (O | label2i)
(This is sometimes called Diracs associativity law.) Hence, the expression
h label1 | O | label2i
is unambiguous.
Remark 5. Please note that
(O | labeli) = hlabel| O
13
4.4. Quantum mechanics: Part II. Observables.

In quantum mechanics, an observable is simply a Hermitian (also
called self-adjoint) operator on a Hilbert space H, i.e., an operator O such
that
O = O .
An eigenvalue a of an operator A is a complex number for which there is

a ket |labeli such that
A |labeli = a |labeli .
The ket |labeli is called an eigenket of A corresponding to the eigenvalue

a.
An important theorem about observables is given below:
Theorem 1. The eigenvalues ai of an observable A are all real numbers.
Moreover, the eigenkets for distinct eigenvalues of an observable are orthogonal.
Definition 1. An eigenvalue is degenerate if there are at least two linearly
independent eigenkets for that eigenvalue. Otherwise, it is non-degenerate
.
Notational Convention: If all the eigenvalues ai of an observable A
are nondegenerate, then we can and do label the eigenkets of A with
the corresponding eigenvalues ai . Thus, we can write:
for each eigenvalue ai .
A |ai i = ai |ai i
Convention: In this paper, unless stated otherwise, we assume that the

eigenvalues of observables are non-degenerate.
One notable exception to the above convention is the measurement
operator
|ai i hai |
for the eigenvalue ai , which is the outer product of ket |ai i with its adjoint
the bra hai |, where we have assumed that |ai i (and hence, hai |) is of unit
length. It has two eigenvalues 0 and 1. 1 is a nondegenerate eigenvalue with
eigenket |ai i. 0 is a degenerate eigenvalue with corresponding eigenkets
{ |aj i }j6=i .
An observable A is said to be complete if its eigenkets |ai i form a basis
of the Hilbert space H. Since by convention all the eigenkets are chosen to
14
be of unit length, it follows that the eigenkets of a complete nondegenerate

observable A form an orthonormal basis of the underlying Hilbert space.
Moreover, given a complete nondegenerate observable A, every ket |i in
H can be written as:
X
|i =
|ai i hai | i
i
Thus, for a complete nondegenerate observable A, we have the following

operator equation which expresses the completeness of A,
X
|ai i hai | = 1
i
In this notation, we also have

A=
X
i
ai |ai i hai | ,
where once again we have assumed that |ai i and hai | are of unit length for
all i.
Example 1. The Pauli spin matrices

0 i
0 1
,
,
2 =
1 =
i
0
1 0
3 =
1
0
0 1
are examples of observables that frequently appear in quantum mechanics

and quantum computation. Their eigenvalues and eigenkets are given in
the following table:
Pauli Matrices
1 =
0 1
1 0
2 =
0 i
i
0
3 =
1
0
0 1
Eigenvalue/Eigenket

1
|0i+|1i
1
+1
= 2
2
1
1
|0i|1i
-1
= 12
2
1

1
|0i+i|1i
1
+1
= 2
2
i
1
|0ii|1i
-1
= 12
2
i

1
+1 |0i =
0
0
-1 |1i =
1
15
4.5. Quantum mechanics: Part III. Quantum measurement General principles.

In this section, A will denote a complete nondegenerate observable with
eigenvalues ai and eigenkets |ai i. We will, on occasion, refer to {|ai i} as
the frame (or the basis) of the observable A.
According to quantum measurement theory, the measurement of an
observable A of a quantum system Q in the state |i produces the eigenvalue ai as the measured result with probability
P rob (Value
ai
is observed) = khai | ik2 ,
and forces the state of the quantum system Q into the state of the corresponding eigenket |ai i.
Since quantum measurement is such a hotly debated topic among physicists, we (in self-defense) quote P.A.M. Dirac[25]:
A measurement always causes the (quantum mechanical) system
to jump into an eigenstate of the dynamical variable that is being
measured.
Thus, the result of the above mentioned measurement of observable A of
a quantum system Q which is in the state |i before the measurement can
be diagrammatically represented as follows:
|i =
|ai i hai | i
First
Meas. of A
=
P rob = khaj | ik2
aj |aj i
|aj i
Second
Meas. of A
=
P rob = 1
Please note that the measured value is the eigenvalue aj with probability
k h aj | i k2 . If the same measurement is repeated on the quantum system
Q after the first measurement, then the result of the second measurement is
no longer stochastic. It produces the previous measured value aj and the
state of Q remains the same, i.e., |aj i .
The observable
|ai i hai |
is frequently called a selective measurement operator (or a filtration)
for ai . As mentioned earlier, it has two eigenvalues 0 and 1. 1 is a nondegenerate eigenvalue with eigenket |aj i, and 0 is a degenerate eigenvalue with
eigenkets {|aj i}j6=i .
|aj i
16
Thus,
|i
Meas. of |ai i hai |

=
P rob = khai | ik2
1 |ai i = |ai i ,
but for j 6= i,
|i
Meas. of |ai i hai |

=
P rob = khaj | ik2
0 |aj i = 0
The above description of quantum measurement is not the most general

possible. For the more advanced quantum measurement theory of probabilistic operator valued measures (POVMs) (a.k.a., positive operator valued measures), please refer to such books as for example [43] and
[72].
4.6. Polarized light: Part III.
surement.
Three examples of quantum mea-
We can now apply the above general principles of quantum measurement

to polarized light. Three examples are given below:4
Example 2.
Vertical
Polaroid
filter
Rt. Circularly
polarized photon
| i =
1
2
P rob =
=
1
2
=
(|li + i |i)
Measurement op.
|li hl|
=
P rob =
1
2
Example 3. A vertically polarized filter followed by a horizontally polarized

filter.
4
Vertically
polarized
photon
|li
The last two examples can easily be verified experimentally with at most three pair
of polarized sunglasses.
0
No photon
Vert.
polar.
filter
photon
|li + |i
17
Vert.
polar.
photon =
P rob = kk2
Horiz.
polar.
filter
No
photon
P rob = 1
=
Normalized so that
kk2 + kk2 = 1
|li
|i h|
|li hl|
Example 4. But if we insert a diagonally polarized filter (by 45o off the
vertical) between the two polarized filters in the above example, we have:
kk2
|li =
1
2
1
2
(|%i + |-i)
|li hl|
|%i =
1
2
1
2
(|li + |i)
|i
|i h|
|%i h%|
where the input to the first filter is |li + |i.

4.7. A Rosetta stone for Dirac notation: Part III. Expected values.
The average value (expected value) of a measurement of an observable
A on a state |i is:
hAi = h| A |i
For, since
X
i
we have
X
hAi = h| A |i = h|
But on the other hand,
|ai i hai | = 1 ,
X
X
|ai i hai | A
|aj i haj | |i =
h | ai i hai | A |aj i haj | i
!
hai | A |aj i = aj hai | aj i = ai ij
i,j
18
Thus,
hAi =
X
i
h | ai i ai hai | i =
X
i
ai khai | ik2
Hence, we have the standard expected value formula,

X
hAi =
ai P rob (Observing ai on input |i) .
i
4.8. Quantum Mechanics: Part IV. The Heisenberg uncertainty

principle.
There is, surprisingly enough, a limitation of what we can observe in the
quantum world.
From classical probability theory, we know that one yardstick of uncertainty is the standard deviation, which measures the average fluctuation
about the mean. Thus, the uncertainty involved in the measurement of a
quantum observable A is defined as the standard deviation of the observed
eigenvalues. This standard deviation is given by the expression
rD
E
(4A)2
U ncertainty(A) =
where
4A = A hAi
Two observables A and B are said to be compatible if they commute,

i.e., if
AB = BA.
Otherwise, they are said to be incompatible.
Let [A, B], called the commutator of A and B, denote the expression
[A, B] = AB BA
In this notation, two operators A and B are compatible if and only if [A, B] =
0.
The following principle is one expression of how quantum mechanics places

limits on what can be observed:
19
Heisenbergs Uncertainty Principle5

D
ED
E 1
(4A)2 (4B)2 |h[A, B]i|2
4
Thus, if A and B are incompatible, i.e., do not commute, then, by measuring A more precisely, we are forced to measure B less precisely, and vice
versa. We can not simultaneously measure both A and B to unlimited
precision. Measurement of A somehow has an impact on the measurement
of B, and vice versa.
4.9. Quantum mechanics: Part V. Dynamics of closed quantum
systems: Unitary transformations, the Hamiltonian, and Schr
odingers
equation.
An operator U on a Hilbert space H is unitary
if
U = U 1 .
Unitary operators are of central importance in quantum mechanics for many

reasons. We list below only two:
Closed quantum mechanical systems transform only via unitary transformations
Unitary transformations preserve quantum probabilities
Let |(t)i denote the state as a function of time t of a closed quantum
mechanical system Q . Then the dynamical behavior of the state of Q is
determined by the Schr
odinger equation
i} |(t)i = H |(t)i ,
t
where } denotes Plancks constant divided by 2, and where H denotes an
observable of Q called the Hamiltonian. The Hamiltonian is the quantum
mechanical analog of the Hamiltonian of classical mechanics. In classical
physics, the Hamiltonian is the total energy of the system.
4.10. The mathematical perspective.
From the mathematical perspective, Schr
odingers equation is written as:
i
U (t) = H(t)U (t) ,
t
}
where
|(t)i = U |(0)i ,
5
We have assumed units have been chosen such that } = 1.
20
and where }i H(t) is a skew-Hermitian operator lying in the Lie algebra of

the unitary group. The solution is given by a multiplicative integral, called
the path-ordered integral,
i
U (t) =
e } H(t)dt,
which is taken over the path }i H(t) in the Lie algebra of the unitary group.
The path-ordered integral is given by:
i
e } H(t)dt = lim
0
0
Y
e } H(k n ) n
k=n
h i
i
t
i
t
i
t
i
t
= lim e } H(n n ) e } H((n1) n ) e } H(1 n ) e } H(0 n )
n
Remark 6. The standard notation for the above path-ordered integral is
Zt
i
P exp
H(t)dt
}
0
If the Hamiltonian H(t) = H is independent of time, then all matrices

commute and the above path-ordered integral simplifies to
i
e } Hdt = e
0
Rt
0
}i Hdt
= e } Ht
Thus, in this case, U (t) is nothing more than a one parameter subgroup of
the unitary group.
5. The Density Operator

5.1. Introducing the density operator.
John von Neumann suggested yet another way of representing the state
of a quantum system.
Let |i be a unit length ket (i.e., h | i = 1) in the Hilbert space H
representing the state of a quantum system6. The density operator
associated with the state ket |i is defined as the outer product of the ket
6
Please recall that each of the kets in the set { |i | C, 6= 0 } represent the same
state of a quantum system. Hence, we can always (and usually do) represent the state of
a quantum system as a unit normal ket, i.e., as a ket such that h | i = 1 .
21
|i (which can be thought of as a column vector) with the bra h| (which

can be thought of as a row vector), i.e.,
= |i h|
The density operator formalism has a number of advantages over the ket
state formalism. One advantage is that the density operator can also be
used to represent hybrid quantum/classical states, i.e., states which are a
classical statistical mixture of quantum states. Such hybrid states may also
be thought of as quantum states for which we have incomplete information.
For example, consider a quantum system which is in the states (each of
unit length)
with probabilities
|1 i , |2i , . . . , |n i
p1, p2, . . . , pn
respectively, where
p1 + p2 + . . . + pn = 1
(Please note that the states |1 i , |2i , . . . , |n i need not be orthogonal.)
Then the density operator representation of this state is defined as
= p1 |1i h1| + p2 |2 i h2| + . . . + pn |n i hn |
If a density operator can be written in the form
= |i h| ,
it is said to represent a pure ensemble . Otherwise, it is said to represent

a mixed ensemble .
5.2. Properties of density operators.
It can be shown that all density operators are positive semi-definite Hermitian operators of trace 1, and vice versa. As a result, we have the following
crisp mathematical definition:
Definition 2. An linear operator on a Hilbert space H is a density operator if it is a positive semi-definite Hermitian operator of trace 1.
It can be shown that a density operator represents a pure ensemble if
and only if 2 = , or equivalently, if and only if T race(2) = 1. For all
ensembles, both pure and mixed, T race(2) 1.
From standard theorems in linear algebra, we know that, for every density
operator , there exists a unitary matrix U which diagonalizes , i.e., such
that U U is a diagonal matrix. The diagonal entries in this matrix are,
22
of course, the eigenvalues of . These are non-negative real numbers which

all sum to 1.
Finally, if we let D denote the set of all density operators for a Hilbert
space H, then iD is a convex subset of the Lie algebra of the unitary group
associated with H.
5.3. Quantum measurement in terms of density operators.
Let {ai } denote the set of distinct eigenvalues ai of an observable A.
Let Pai denote the projection operator that projects the underlying Hilbert
space onto the eigenspace determined by the eigenvalue ai . For example, if
ai is a non-degenerate eigenvalue, then
Pai = |ai i hai |
Finally, let Q be a quantum system with state given by the density operator
.
If the quantum system Q is measured with respect to the observable A,

then with probability
pi = T race (Pai )
the resulting measured eigenvalue is ai , and the resulting state of Q is given
by the density operator
i =
Pai Pai
.
T race (Pai )
Moreover, for an observable A, the averaged observed eigenvalue expressed

in terms of the density operator is:
hAi = trace(A)
Thus, we have extended the definition of hAi so that it applies to mixed
as well as pure ensembles, i.e., generalized the following formula to mixed
ensembles:
hAi = h | A | i = trace (|i h| A) = trace(A) .
23
5.4. Some examples of density operators.

For example, consider the following mixed ensemble of the polarization
state of a photon:
Example 5.
Ket
|li |%i
Prob.
3
4
1
4
In terms of the basis |i, |li of the two dimensional Hilbert space H, the
density operator of the above mixed ensemble can be written as:
=
3
4
3
4
3
4
|li hl| +
1
0
1
4
1 0
0 0
|%i h%|
1 0

1
8
1
4
1 1
1 1

1/2
1/ 2

1/ 2 1/ 2
7
8
1
8
1
8
1
8
Example 6. The following two preparations produce mixed ensembles with

the same density operator:
Ket
|li |i
Prob.
1
2
and
1
2
Ket
|%i |-i
Prob.
1
2
1
2
For, for the left preparation, we have
1
2
1
2
1
2
|li hl| +

1
0
1 0
0 1
1
2
|i h|
1 0

And for the right preparation, we have
1
2
0
1
0 1
24
1
2
|%i h%| + 12 |-i h-|
1 1
2 2
1
4
1
1
1 1
1 1
1
2
1 1
1
4
1 1
2 2
1 1
1
1
1
1
1
2
1
2
1 0
0 1
1 1

There is no way of physically distinquishing the above two mixed ensembles

which were prepared in two entirely different ways. For the density operator
represents all that can be known about the state of the quantum system.
5.5. The partial trace of a linear operator.

In order to deal with a quantum system composed of many quantum
subsystems, we need to define the partial trace.
Let
O : H H HomC (H, H)
be a linear operator on the Hilbert space H.
Since Hilbert spaces are free algebraic objects, it follows from standard
results in abstract algebra7 that
HomC (H, H)
= H H ,
where we recall that
H = HomC (H, C) .
Hence, such an operator O can be written in the form

X
O=
a |h i hk| ,
where the kets |h i lie in H and the bras hk | lie in H .

Thus, the standard trace of a linear operator
T race : HomC (H, H) C
7
See for example [55].
25
is nothing more than a contraction, i.e.,

X
T race(O) =
a h k | h i ,
i.e., a replacement of each outer product |h i hk | by the corresponding

bracket h k | h i.
We can generalize the T race as follows:
Let H now be the tensor product of Hilbert spaces H1 , H2 , . . . ,Hn , i.e.,
H=
n
O
Hj .
j=1
Then it follows once again from standard results in abstract algebra that
HomC (H, H)
=
n
O
j=1
Hj Hj
Hence, the operator O can be written in the form

O=
n
O
j=1
|h,j i hk,j | ,
where, for each j, the kets |h,j i lie in Hj and the bras hk,j | lie in Hj for
all .
Next we note that for every subset I of the set of indices J = {1, 2, . . . , n},
we can define the partial trace over I, written
O
O
O
O
Hj ,
Hj
Hj ,
Hj HomC
T raceI : HomC
jJ
jJ
jJ I
jJ I
as the contraction on the indices I, i.e.,
X
Y
O
T raceI (O) =
a
h k,j | h,j i
| h,j i h k,j | .
jI
jJ I
For example, let H1 and H0 be two dimensional Hilbert spaces with selected orthonormal bases {|01 i , |11i} and {|00i , |10 i}, respectively. Thus,
{|01 00i , |01 10i , |11 00i , |11 10i} is an orthonormal basis of H = H1 H0 .
26
Let HomC (H, H) be the operator

|0100 i |1110 i
h01 00| h11 10|
2
2
1
= (|0100i h01 00| |01 00i h1110 | |1110i h01 00| + |11 10i h1110 |)
2
which in terms of the basis {|01 00i , |01 10i , |11 00i , |11 10i} can be written as
the matrix
1 0 0 1
1 0 0 0
0
,
=
0
2 0 0 0
1 0 0
1
where the rows and columns are listed in the order |0100i, |0110 i, |1100 i,
|11 10i
The partial trace T race0 with respect to I = {0} of is
1 = T race0 ()
1
= T race0 (|01 00i h01 00| |0100 i h1110 | |11 10i h01 00| + |1110 i h1110 |)
2
1
= (h00 | 00i |01i h01 | h10 | 00i |01i h11 | h00 | 10i |11 i h01 | + h10 | 10i |11 i h11|)
2
1
= (|01 i h01| |01i h11 |)
2
which in terms of the basis {|01i , |11 i} becomes

1 1 0
1 = T race0() =
,
2 0 1
where the rows and columns are listed in the order |01 i, |11i .
5.6. Multipartite quantum systems.

One advantage density operators have over kets is that they provide us
with a means for dealing with multipartite quantum systems.
Definition 3. Let Q1 , Q2 , . . . , Qn be quantum systems with underlying
Hilbert spaces H1 , H2 , . . . , Hn , respectively. The global quantum system Q
consisting of the quantum systems Q1 , Q2 , . . . , Qn is called a multipartite
quantum system. Each of the quantum systems Qj (j = 1, 2, . . . , n) is
called a constituent part of Q . The underlying Hilbert space H of Q
is the tensor product of the Hilbert spaces of the constituent parts, i.e.,
H=
n
O
j=1
Hj .
27
If the density operator is the state of a multipartite quantum system

Q, then the state of each constituent part Qj is the density operator j
given by the partial trace
j = T raceJ {j} () ,
where J = {1, 2, . . . , n} is the set of indices.
Obviously, much more can be said about the states of multipartite systems and their constituent parts. However, we will forego that discussion
until after we have had an opportunity introduce the concepts of quantum
entanglement and von Neumann entropy.
5.7. Quantum dynamics in density operator formalism.
Under a unitary transformation U , a density operator transforms according to the rubric:
7 U U
Moreover, in terms of the density operator, Schr
odingers equation8 becomes:
i}
= [H, ] ,
t
where [H, ] denotes the commutator of H and , i.e.,
[H, ] = H H
5.8. The mathematical perspective.
From the mathematical perspective, one works with i instead of because i lies in the Lie algebra of the unitary group. Thus, the density operator transforms under a unitary transformation U according to the rubric:
i 7 AdU (i) ,
where AdU denotes the big adjoint representation.
From the mathematical perspective, Schr
odingers equation is in this case
more informatively written as:
1
(i)
= adiH (i) ,
t
}
8
Schr
odingers equation determines the dynamics of closed quantum systems. However,
non-closed quantum systems are also of importance in quantum computation and quantum
information theory. See for example the Schumachers work on superoperators, e.g., [76].
28
where ad i H denotes the little adjoint representation . Thus, the solu}

tion to the above form of Schr
odingers equation is given by the path ordered
integral:
!
}1 (adiH(t) )dt
e
=
t
where 0 denotes the density operator at time t = 0.
6. The Heisenberg model of quantum mechanics
Consider a computing device with inputs and outputs for which we have
no knowledge of the internal workings of the device. We are allowed to
probe the device with inputs and observe the corresponding outputs. But
we are given no information as to how the device performs its calculation.
We call such a device a blackbox computing device.
For such blackboxes, we say that two theoretical models for blackboxes
are equivalent provided both predict the same input/output behavior. We
may prefer one model over the other for various reasons, such as simplicity,
aesthetics, or whatever meets our fancy. But the basic fact is that each of
the two equivalent models is just as correct as the other.
In like manner, two theoretical models of the quantum world are said to
be equivalent if they both predict the same results in regard to quantum
measurements.
Up to this point, we have been describing the Schr
odinger model of quantum mechanics. However, shortly after Schr
odinger proposed his model
for the quantum world, called the Schr
odinger picture, Heisenberg proposed yet another, called the Heisenberg picture. Both models were
later proven to be equivalent.
In the Heisenberg picture, state kets remain stationary with time, but
observables move with time. While state kets, and hence density operators,
remain fixed with respect to time, the observables A change dynamically as:
A 7 U AU
under a unitary transformation U = U (t), where the unitary transformation
is determined by the equation
i}
U
= HU
t
29
It follows that the equation of motion of observables is according to the

following equation
i}
A
= [A, H]
t
One advantage the Heisenberg picture has over the Schr

odinger picture is
that the equations appearing in it are similar to those found in classical
mechanics.
30
In summary, we have the following table which contrasts the two pictures:
State ket
Density
Operator
Schr
odinger Picture
Moving
Heisenberg Picture
Stationary
|0i 7 |i = U |0i
|0 i
Moving
Stationary
0 7 = U 0U = AdU (0 )
Stationary
Moving
A0
A0 7 A = U A0 U = AdU (A0)
Stationary
Stationary
aj
aj
Observable
Observable
Eigenvalues
Observable
Frame
Stationary
P
A0 = j aj |aj i0 haj |0
Moving
P
A0 = j aj |aj i0 haj |0
At =
aj |aj it haj |t
where |aj it = U |aj i0
Dynamical
Equations
Measurement
(S)
i} U
U
t = H
(H)
i} U
U
t = H
i} t
|i = H (S) |i

(H)
i} A
t = A, H
Measurement of observable A0
produces eigenvalue aj with
probability

haj | |i2 = haj | |i2
0
where
H (H) = U H (S) U
Measurement of observable A
produces eigenvalue aj with
probability

haj | |0i2 = haj | |i2
t
31
It follows that the Schr

odinger Hamiltonian H (S) and the Heisenberg Hamiltonian are related as follows:
H (S)
H (H)
=U
U ,
t
t
U
where terms containing U
t and t have cancelled out as a result of the
Schr
odinger equation.
We should also mention that the Schr
odinger and Heisenberg pictures can
be transformed into one another via the mappings:
S H
H S
(S)

7 (H) = U (S)
(H)

7 (S) = U (H)
A(S) 7 A(H) = U A(S) U
A(H) 7 A(S) = U A(H)U
A(S) 7 A(H) = U A(S) U
A(H) 7 A(S) = U A(H)U
(S) 7 (H) = U (S)U
(H) 7 (S) = U (H)U
Obviously, much more could be said on this topic.

For quantum computation from the perspective of the Heisenberg model,
please refer to the work of Deutsch and Hayden[23], and also to Gottesmans
study of the ancient Hittites :-) [31].
7. Quantum entanglement
7.1. The juxtaposition of two quantum systems.
Let Q1 and Q2 be two quantum systems that have been separately prepared respectively in states |1i and |2 i, and that then have been united
without interacting. Because Q1 and Q2 have been separately prepared
without interacting, their states |1i and |2i respectively lie in distinct
Hilbert spaces H1 and H2 . Moreover, because of the way in which Q1 and
Q2 have been prepared, all physical predictions relating to one of these quantum systems do not depend in any way whatsoever on the other quantum
system.
The global quantum system Q consisting of the two quantum systems Q1
and Q2 as prepared above is called a juxtaposition of the quantum systems
Q1 and Q2 . The state of the global quantum system Q is the tensor product
of the states |1i and |2i. In other words, the state of Q is:
|1i |2i H1 H2
32
7.2. An example: An n-qubit register Q consisting of the juxtaposition of n qubits.

Let H be a two dimensional Hilbert space, and let {|0i , |1i} denote an arbitrarily selected orthonormal basis9. Let Hn1 , Hn2 , . . . ,H0 be distinct
Hilbert spaces, each isomorphic to H, with the obvious induced orthonormal
bases
{ |0n1 i , |1n1 i } , { |0n2 i , |1n2 i } , . . . , { |00i , |10 i }
respectively.
Consider n qubits Qn1 , Qn2 , . . . , Q0 separately prepared in the states
1
1
1
(|0n1 i + |1n1 i) , (|0n2 i + |1n2 i) , . . . , (|00i + |10i) ,
2
2
2
respectively. Let Q denote the global system consisting of the separately
prepared (without interacting) qubits Qn1 , Qn2 , . . . , Q0 . Then the state
|i of Q is:
1
1
1
|i = (|0n1 i + |1n1 i) (|0n2 i + |1n2 i) . . . (|00i + |10i)
2
2
2

1 n
=
(|0n1 0n2 . . . 0100i + |0n1 0n2 . . . 01 10i + . . . + |1n1 1n2 . . . 1110i)
2
which lies in the Hilbert space
H = Hn1 Hn2 . . . H0 .
Notational Convention: We will usually omit subscripts whenever they
can easily be inferred from context.
Thus, the global system Q consisting of the n qubits Qn1 , Qn2 , . . . ,
Q0 is in the state

n1
O
1 n
|i =
(|00 . . . 00i + |00 . . . 01i + . . . + |11 . . . 11i)
H
2
0
The reader should note that the n-qubit register Q is a superposition of
kets with labels consisting of all the binary n-tuples. If each binary n-tuple
bn1 bn2 . . . b0 is identified with the integer
bn1 2n1 + bn2 2n2 + . . . + b020 ,
9
We obviously have chosen to label the basis elements in a suggestive way.
33
i.e., if we interpret each binary n-tuple as the radix 2 representation of an

integer, then we can rewrite the state as

1 n
|i =
(|0i + |1i + |2i + . . . + |2n 1i) .
2
In other words, this n-qubit register contains all the integers from 0 to 2n 1
in superposition. But most importantly, it contains all the integers 0 to
2n 1 simultaneously!
This is an example of the massive parallelism that is possible within quantum computation. However, there is a downside. If we observe (measure)
the register, then all the massive parallelism disappears. On measurement, the quantum world selects for us one and only one of the 2n integers.
probability of observing any particular one of the integers is
The
n 2

1/ 2 = ( 21 )n . The selection of which integer is observed is unfortunately not made by us, but by the quantum world.
Thus, harnessing the massive parallelism of quantum mechanics is no easy
task! As we will see, a more subtle approach is required.
7.3. An example of the dynamic behavior of a 2-qubit register.

We now consider the previous n-qubit register for n = 2.
bases described in the previous section, we have:

0
1
1
|0i = |00i =
0
0
0

1
1
0
|1i = |01i =
0
0
1

0
0
1
|2i = |10i =
1
1
0

0
0
0
|3i = |11i =
0
1
1
In terms of the
34
Let us assume that the initial state |it=0 of our 2-qubit register is
1

|0i |1i
1
1
1
0
|it=0 =
|0i = (|00i |10i) = (|0i |2i) =
1
2
2
2
2
0
Let us also assume that from time t = 0 to time t = 1 the dynamical
behavior of the above 2-qubit register is determined by a constant Hamiltonian H, which when written in terms of the basis {|00i , |01i , |10i , |11i} =
{|0i , |1i , |2i , |3i} is given by
0 0
0
0
}
0
0
,
0 0
H=
0 0
1 1
2
0 0 1
1
where the rows and the columns are listed in the order |00i, |01i, |10i, |11i,
i.e., in the order |0i, |1i, |2i, |3i.
Then, as a consequence of Schr
odingers equation, the Hamiltonian H
determines a unitary transformation
UCN OT =
1
0
=
0
0
e } Hdt = e
0
0
1
0
0
0
0
0
1
R1
0
}i Hdt
= e } H
0
0
= |0i h0| + |1i h1| + |2i h3| + |3i h2|
1
0
which moves the 2-qubit register from the initial state |it=0 at time t = 0
to |it=1 = UCN OT |it=0 at time t = 1. Then
1 0 0 0
1
0 1 0 0 1 0
|it=1 = UCN OT |it=0 =

0 0 0 1 2 1
0 0 1 0
0
1
0
= 1 (|00i |11i) = 1 (|0i |3i)
=
0
2
2
2
1
The resulting state (called an EPR pair of qubits for reasons we shall
later explain) can no longer be written as a tensor product of two states.
Consequently, we no longer have the juxtaposition of two qubits.
35
Somehow, the resulting two qubits have in some sense lost their separate
identities. Measurement of any one of the qubits immediately impacts the
other.
For example, if we measure the 0-th qubit (i.e., the right-most qubit), the
EPR state in some sense jumps to one of two possible states. Each of the
two possibilities occurs with probability 21 , as indicated in the table below:
1
2
(|0100 i |1110 i)
...
Meas.
0-th
Qubit
P rob = 21
|0100 i
&&&
P rob = 21
|11 10i
Thus we see that a measurement of one of the qubits causes a change in the
other.
7.4. Definition of quantum entanglement.

The above mentioned phenomenon is so unusual and so non-classical that
it warrants a name.
Definition 4. Let Q1 , Q2 , . . . , Qn be quantum systems with underlying
Hilbert spaces H1 , H2 , . . . , Hn , respectively. Then the global quantum
system Q consisting of the quantum
Nsystems Q1, Q2 , . . . , Qn is said to be
entangled if its state |i H = nj=1 Hj can not be written in the form
|i =
n
O
j=1
|j i ,
where each ket |j i lies in the Hilbert space Hj for, j = 1, 2, . . . , n.

also say that such a state |i is entangled.
Thus, the state
1
|it=1 = (|00i |11i)
2
of the 2-qubit register of the previous section is entangled.
We
36
Remark 7. In terms of density operator formalism, a pure ensemble is

entangled if it can not be written in the form
=
n
O
j ,
j=1
where the j s denote density operators.

Please note that we have defined entanglement only for pure ensembles.
For mixed ensembles, entanglement is not well understood10 . As a result,
the right definition of entanglement of mixed ensembles is still unresolved.
We give one definition below:
Definition 5. A density operator on a Hilbert space H is said to be entangled with respect to the Hilbert space decomposition
H=
if it can not be written in the form
=
`
X
k=1
n
O
Hj
j=1
n
O
k
(j,k) ,
j=1
for some positive integer `, where the k s are positive real numbers such
that
`
X
k = 1 .
k=1
and where each (j,k) is a density operator on the Hilbert space Hj.
Readers interested in pursuing this topic further should refer to the works
of Bennett , the Horodeckis, Nielsen, Smolin, Wootters , and others[6], [45],
[59], [67].
7.5. Einstein, Podolsky, Rosens (EPRs) grand challenge to quantum mechanics.
Albert Einstein was skeptical of quantum mechanics, so skeptical that he
together with Podolsky and Rosen wrote a joint paper[26] appearing in 1935
challenging the very foundations of quantum mechanics. Their paper hit
the scientific community like a bombshell. For it delivered a direct frontal
attack at the very heart and center of quantum mechanics.
10
Quantum entanglement is not even well understood for pure ensembles.
37
At the core of their objection was quantum entanglement. Einstein

and his colleagues had insightfully recognized the central importance of this
quantum phenomenon.
Their argument centered around the fact that quantum mechanics violated either the principle of non-locality11 or the principle of reality12 .
They argued that, as a result, quantum mechanics must be incomplete, and
that quantum entanglement could be explained by missing hidden variables.
For many years, no one was able to conceive of an experiment that could
determine which of the two theories, i.e., quantum mechanics or EPRs
hidden variable theory, was correct. In fact, many believed that the two
theories were not distinguishable on physical grounds.
It was not until Bell developed his famous inequalities [3],[4], [13], that a
physical criterion was found to distinquish the two theories. Bell developed
inequalities which, if violated, would clearly prove that quantum mechanics
is correct, and hidden variable theories are not. Many experiments were
performed. Each emphatically supported quantum mechanics, and clearly
demonstrated the incorrectness of hidden variable theory. Quantum mechanics was the victor!
7.6. Why did Einstein, Podolsky, Rosen (EPR) object?
But why did Einstein and his colleagues object so vehemently to quantum
entanglement?
As a preamble to our answer to this question, we note that Einstein and
his colleagues were convinced of the validity of the following two physical
orinciples:
1) The principle of local interactions , i.e., that all the known forces
of nature are local interactions,
2) The principle of non-locality, i.e., that spacelike separated regions
of spacetime are physically independent of one another.
Their conviction in regard to principle 1) was based on the fact that all
four known forces of nature, i.e., gravitational, electromagnetic, weak, and
strong forces, are local interactions. By this we mean:
i) They are mediated by another entity, e.g., graviton, photon, etc.
ii) They propagate no faster than the speed c of light
iii) Their strength drops off with distance
11
We will later explain the principle of non-locality.

For an explanation of the principle of reality as well as the principle of non-localty,
please refer, for example, to [72], [13].
12
38
Their conviction in regard to principle 2) was based on the following

reasoning:
Two points in spacetime P1 = (x1 , y1, z1, t1) and P2 = (x2 , y2, z2, t2 )
are separated by a spacelike distance provided the distance between
(x1 , y1, z1) and (x2 , y2, z2) is greater than c |t2 t1 |, i.e.,
Distance ((x1 , y1, z1) , (x2 , y2, z2)) > c |t2 t1 | ,
where c denotes the speed of light. In other words, no signal can travel
between points that are said to be separated by a spacelike distance unless
the signal travels faster than the speed of light. But because of the basic
principles of relativity, such superluminal communication is not possible.
Hence we have:
The principle of non-locality: Spacelike separated regions of spacetime are physically independent. In other words, spacelike separated
regions can not influence one another.
7.6.1. EPRs objection.
We now are ready to explain why Einstein and his colleagues objected
so vehemently to quantum entanglement. We explain Bohms simplified
version of their argument.
Consider a two qubit quantum system that has been prepared by Alice
in her laboratory in the state
1
|i = (|01 00i |1110 i) .
2
After the preparation, she decides to keep qubit #1 in her laboratory, but
enlists Captain James T. Kirk of the Starship Enterprise to transport qubit
#0 to her friend Bob 14 who is at some far removed distant part of the
universe, such as at a Federation outpost orbiting about the double star
Alpha Centauri in the constellation Centaurus.
13
After Captain Kirk has delivered qubit #0, Alices two qubits are now
separated by a spacelike distance. Qubit #1 is located in her Earth based
laboratory. Qubits #0 is located with Bob at a Federation outpost orbiting
Alpha Centauri. But the two qubits are still entangled, even in spite of the
fact that they are separated by a spacelike distance.
If Alice now measures qubit #1 (which is located in her Earth based
laboratory), then the principles of quantum mechanics force her to conclude
that instantly, without any time lapse, both qubits are effected. As a
13
Alice is a well known personality in quantum computation, quantum cryptography,

and quantum information theory.
14
Bob is another well known personality in quantum computation, quantum cryptography, and quantum information theory.
39
result of the measurement, both qubits will be either in the state |0100 i or
the state |11 10i, each possibility occurring with probability 1/2.
This is a non-local interaction. For,
The interaction occurred without the presence of any force. It was
not mediated by anything.
The measurement produced an instantaneous change, which was certainly faster than the speed of light.
The strength of the effect of the measurement did not drop off with
distance.
No wonder Einstein was highly skeptical of quantum entanglement. Yet
puzzlingly enough, since no information is exchanged by the process, the
principles of general relativity are not violated. As a result, such an effect
can not be used for superluminal communication.
For a more in-depth discussion of the EPR paradox and the foundations
of quantum mechanics, the reader should refer to [13].
7.7. Quantum entanglement: The Lie group perspective.
Many aspects of quantum entanglement can naturally be captured in
terms of Lie groups and their Lie algebras.
Let
H = Hn1 Hn2 . . . H0 =
n1
Hj
be a decomposition of a Hilbert space H into the tensor product of the

Hilbert spaces Hn1 , Hn2 , . . . ,H0 . Let U = U(H), Un1 = U(Hn1 ),
Un2 = U(Hn2 ), . . . ,U0 = U(H0 ), denote respectively the Lie groups of
all unitary transformations on H, Hn1 , Hn2 , . . . ,H0 . Moreover, let
u = u(H), un1 = un1 (Hn1 ), un2 = un2 (Hn2 ), . . . , u0 = u0 (H0 )
denote the corresponding Lie algebras.
Definition 6. The local subgroup L = L(H) of U = U(H) is defined as
the subgroup
O
L = Un1 Un2 . . . U0 = n1
Uj .
0
The elements of L are called local unitary transformations . Unitary

transformations which are in U but not in L are called global unitary
transformations. The corresponding lie algebra
` = un1 un2 . . . u0
40
is called the local Lie algebra, where denotes the Kronecker sum15.
Local unitary transformations can not entangle quantum systems with

respect to the above tensor product decomposition. However, global unitary
transformations are those unitary transformations which can and often do
produce interactions which entangle quantum systems. This leads to the
following definition:
Definition 7. Two states |1 i and |2i in H are said to be locally equivalent ( or, of the same entanglement type) , written
|1i |2i ,
local
if there exists a local unitary transformation U L such that

U |1i = |2i .
The equivalence classes of local equivalence are called the entanglelocal
ment classes of H. Two density operators 1 and 2, (and hence, the

corresponding two skew Hermitian operators i1 and i2 lying in u) are said
to be locally equivalent ( or, of the same entanglement type), written
1 2 ,
local
if there exists a local unitary transformation U L such that

AdU (1) = 2 ,
where AdU denotes the big adjoint representation, i.e., AdU (i) = U (i)U .
The equivalence classes under this relation are called entanglement classes
of the Lie algebra u(H).
Thus, the entanglement classes of the Hilbert space H are just the orbits
of the group action of L(H) on H. In like manner, the entanglement classes
of the Lie algebra u(H) are the orbits of the big adjoint action of L(H) on
u(H). Two states are entangled in the same way if and only if they lie in
the same entanglement class, i.e., the same orbit.
For example, let us assume that Alice and Bob collectively possess two
qubits QAB which are in the entangled state

1
1
|0B 0A i + |1B 1A i
0
,
=
|1 i =
0
2
2
1
15
The Kronecker sum A
B is defined as
A
B = A1+1B ,
where 1 denotes the identity transformation.
41
and moreover that Alice possesses qubit labeled A, but not the qubit labeled
B, and that Bob holds qubit B, but not qubit A. Let us also assume that
Alice and Bob are also separated by a spacelike distance. As a result, they
can only apply local unitary transformations to the qubits that they possess.
Alice could, for example, apply the local unitary transformation
0
0 1 0

0
0 0 1
1 0
0 1
UA =
1
0 0 0
0 1
1 0
0 1 0 0
to her qubit to move Alices and Bobs qubits A and B respectively into the
state
0
|0B 1A i |1B 0A i
1 1
,
|2i =
=
2
2 1
0
Bob also could accomplish the same by applying the local unitary transformation
0
1
0
0

1
1 0
0 1
0 0
0
UB =
=
0
0 1
1
0
0 0 1
0
0 1
0
to his qubit.
By local unitary transformations, Alice and Bob can move the state of
their two qubits to any other state within the same entanglement class.
But with local unitary transformations, there is no way whatsoever that
Alice and Bob can transform the two qubits into a state lying in a different
entanglement class (i.e., a different orbit), such as
|3i = |0B 0A i .
The only way Alice and Bob could transform the two qubits from state
|1 i to the state |3i is for Alice and Bob to come together, and make the
two qubits interact with one another via a global unitary transformation
such as
1
0 0 1
1 0
1 1 0
UAB =
2 0 1 1 0
1
0 0 1
The main objective of this approach to quantum entanglement is to determine when two states lie in the same orbit or in different orbits? In other
words, what is needed is a complete set of invariants, i.e., invariants that
42
completely specify all the orbits ( i.e., all the entanglement classes).
save this topic for another lecture[61].
We
At first it would seem that state kets are a much better vehicle than
density operators for the study of quantum entanglement. After all, state
kets are much simpler mathematical objects. So why should one deal with
the additional mathematical luggage of density operators?
Actually, density operators have a number of advantages over state kets.
The most obvious advantage is that density operators certainly have an upper hand over state kets when dealing with mixed ensembles. But their
most important advantage is that the orbits of the adjoint action are actually manifolds, which have a very rich and pliable mathematical structure.
Needless to say, this topic is beyond the scope of this paper.
Remark 8. It should also be mentioned that the mathematical approach
discussed in this section by no means captures every aspect of the physical
phenomenon of quantum entanglement. The use of ancilla and of classical
communication have not been considered. For an in-depth study of the relation between quantum entanglement and classical communication (including
catalysis), please refer to the work of Jonathan, Nielson, and others[67].
In regard to describing the locality of unitary operations, we will later have
need for a little less precision than that given above in the above definitions.
So we give the following (unfortunately rather technical) definitions:
Definition 8. Let H, Hn1 , Hn2 , . . . ,H0 be as stated above. Let P =
{B } be a partition of the set of indices {0, 1, 2, . . . , n 1}, i.e., P is a
collection
of disjoint subsets B of {0, 1, 2, . . . , n 1}, called blocks, such
S
that B = {0, 1, 2, . . . , n 1}. Then the P-tensor product decomposition of H is defined as
O
H=
HB ,
where
B P
HB =
jB
Hj ,
for each block B in P. Also the subgroup of P-local unitary transformations LP (H) is defined as the subgroup of local unitary transformations
of H corresponding to the P-tensor decomposition of H.
We define the fineness of a partition P, written f ineness(P), as the
maximum number of indices in a block of P. We say that a unitary transformation U of H is sufficiently local if there exists a partition P with sufficiently small f ineness(P) (e.g., f ineness(P) 3) such that U LP (H).
43
Remark 9. The above lack of precision is needed because there is no way to

know what kind (if any) of quantum computing devices will be implemented
in the future. Perhaps we will at some future date be able to construct
quantum computing devices that locally manipulate more than 2 or 3 qubits
at a time?
8. Entropy and quantum mechanics
8.1. Classical entropy, i.e., Shannon Entropy.
Let S be a probability distribution on a finite set {s1 , s2, . . . , sn } of elements called symbols given by
Prob (sj ) = pj ,
P
where nj=1 pj = 1. Let s denote the random variable (i.e., finite memoryless stochastic source) that produces the value sj with probability
pj .
Definition 9. The classical entropy (also called the Shannon entropy)

H(S) of a probability distribution S (or of the source s) is defined as:
n
X
H(S) = H(s) = pj lg(pj ) ,
j=1
where lg denotes the log to the base 2 .
Classical entropy H(S) is a measure of the uncertainty inherent in the

probability distribution S. Or in other words, it is the measure of the
uncertainty of an observer before the source s outputs a symbol sj .
One property of such classical stochastic sources we often take for
granted is that the output symbols sj are completely distinguishable from
one another. We will see that this is not necessarily the case in the strange
world of the quantum.
44
8.2. Quantum entropy, i.e., Von Neumann entropy.

Let Q be a quantum system with state given by the desity operator .
Then there are many preparations
Preparation
|1i
p1
|2 i
p2
...
...
|n i
pn
which will produce the same state . These preparations are classical stochastic sources with classical entropy given by
X
H = pj lg(pj ) .
Unfortunately, the classical entropy H of a preparation does not necessarily

reflect the uncertainty in the resulting state . For two different preparations P1 and P2, having different entropies H (P1 ) and H (P2), can (and
often do) produce the same state . The problem is that the states of the
preparation my not be completely physically distinguishable from one another. This happens when the states of the preparation are not orthogonal.
(Please refer to the Heisenberg uncertainty principle.)
John von Neumann found that the true measure of quantum entropy can
be defined as follows:
Definition 10. Let Q be a quantum system with state given by the density
operator . Then the quantum entropy (also called the von Neumann
entropy) of Q, written S(Q), is defined as
S(Q) = T race ( lg ) ,
where lg denotes the log to the base 2 of the operator .

Remark 10. The operator lg exists and is an analytic map 7 lg
given by the power series
( I)n
1 X
(1)n+1
lg =
ln 2 n=1
n
provided that is sufficiently close to the identity operator I, i.e., provided

where
k Ik < 1 ,
kAvk
.
vH kvk
kAk = sup
It can be shown that this is the case for all positive definite Hermitian operators of trace 1.
For Hermitian operators of trace 1 which are not positive definite, but
only positive semi-definite (i.e., which have a zero eigenvalue), the logarithm
45
lg() does not exist. However, there exists a sequence 1 , 2, 3, . . . of positive definite Hermitian operators of trace 1 which converges to , i.e., such
that
= lim k
k
It can then be shown that the limit

lim k lg k
exists.
Hence, S() is defined and exists for all density operators .
Quantum entropy is a measure of the uncertainty at the quantum level.
As we shall see, it is very different from the classical entropy that arises
when a measurement is made.
One important feature of quantum entropy S() is that it is invariant
under the adjoint action of unitary transformations, i.e.,

S ( AdU () ) = S U U = S() .
It follows that, for closed quantum systems, it is a dynamical invariant.

As the state moves according to Schr
odingers equation, the quantum
entropy S() of remains constant. It does not change unless measurement
is made, or, as we shall see, unless we ignore part of the quantum system.
Because of unitary invariance, the quantum entropy can be most easily
computed by first diagonalizing with a unitary transformation U , i.e.,
U U = 4( ) ,
where 4( ) denotes the diagonal matrix with diagonal = (1, 2, . . . , n).

Once has been diagonalized , we have

S() = T race 4( ) lg 4( )
= T race ( 4(1 lg 1, 2 lg 2, . . . , n lg n ) )
n
X
= j lg j ,
j=1
where the j s are the eigenvalues of , and where 0 lg 0 0.
Please note that, because is positive semi-definite Hermitian of trace 1,

all the eigenvalues of are non-negative real numbers such that
n
X
j = 1 .
j=1
46
As an immediate corollary we have that the quantum entropy of a pure

ensemble must be zero, i.e.,
pure ensemble = S() = 0
There is no quantum uncertainty in a pure ensemble. However, as expected,
there is quantum uncertainty in mixed ensembles.
8.3. How is quantum entropy related to classical entropy?
But how is classical entropy H related to quantum entropy S?
Let A be an observable of the quantum system Q. Then a measurement
of A of Q produces an eigenvalue ai with probability
pi = T race (Pai ) ,
where Pai denotes the projection operator for the eigenspace of the eigenvalue ai . For example, if ai is a non-degenerate eigenvalue, then Pai =
|ai i hai | .
In other words, measurement of A of the quantum system Q in state
can be identified with a classical stochastic source with the eigenvalues ai
as output symbols occurring with probability pi . We denote this classical
stochastic source simply by (, A) .
The two entropies S() and H(, A) are by no means the same. One is a
measure of quantum uncertainty before measurement, the other a measure
of the classical uncertainty that results from measurement. The quantum
entropy S() is usually a lower bound for the classical entropy, i.e.,
S() H(, A) .
If A is a complete observable (hence, non-degenerate), and if A is compatible
with , i.e., [, A] = 0, then S() = H(, A).
8.4. When a part is greater than the whole Ignorance = uncertainty.
Let Q be a multipartite quantum system with constituent parts Qn1 ,
. . . ,Q1 , Q0 , and let the density operator denote the state of Q. Then
from section 5.6 of this paper we know that the state j of each constituent
47
part Qj is given by the partial trace over all degrees of freedom except
Qj , i.e., by
j =
T race ()
.
0 k n1
k 6= j
By applying the above partial trace, we are focusing only on the quantum
system Qj , and literally ignoring the remaining constituent parts of Q.
By taking the partial trace, we have done nothing physical to the quantum
system. We have simply ignored parts of the quantum system.
What is surprising is that, by intentionally ignoring part of the quantum system, we can in some cases create more quantum uncertainty. This
happens when the constituent parts of Q are quantum entangled.
For example, let Q denote the bipartite quantum system consisting of two
qubits Q1 and Q0 in the entangled state
|01 00i |11 10i
.
|Q i =
2
The corresponding density operator Q is
1
Q = (|01 00i h0100 | |0100 i h1110| |11 10i h01 00| + |1110 i h1110|)
2
1 0 0 1
1 0 0 0
0
0 0 0
0
2
1 0 0
1
Since Q is a pure ensemble, there is no quantum uncertainty, i.e.,
S (Q ) = 0 .
Let us now focus on qubit #0 (i.e., Q0). The resulting density operator
0 for qubit #0 is obtained by tracing over Q1 , i.e.,

1
1 1 0
0 = T race1 (Q) = ( |0i h0| + |1i h1| ) =
.
2
2 0 1
Hence, the quantum uncertainty of qubit #0 is

S(0) = 1 .
Something most unusual, and non-classical, has happened. Simply by

ignoring part of the quantum system, we have increased the quantum uncertainty. The quantum uncertainty of the constituent part Q0 is greater
48
than that of he whole quantum system Q. This is not possible in the classical world, i.e., not possible for Shannon entropy. (For more details, see
[15].)
9. There is much more to quantum mechanics
There is much more to quantum mechanics. For more in-depth overviews,

there are many outstanding books. Among such books are [13], [16], [25],
[28], [40], [41], [43], [46], [48], [64], [68], [70], [72], [74], [75], and many more.
Part 3.
Part of a Rosetta Stone for Quantum

Computation
10. The Beginnings of Quantum Computation - Elementary

Quantum Computing Devices
We begin this section with some examples of quantum computing devices.

By a quantum computing device16 we mean a unitary transformation U
that is the composition of finitely many sufficiently local unitary transformations, i.e.,
U = Un1 Un2 . . . U1 U0 ,
where Un1 , Un2 , . . . ,U1 ,U0 are sufficiently local17 unitary transformations. Each Uj is called a computational step of the device.
Our first examples will be obtained by embedding classical computing
devices within the realm of quantum mechanics. We will then look at some
other quantum computing devices that are not the embeddings of classical
devices.
16
Unfortunately, Physicists have stolen the akronym QCD. :-)

See Definition 8 in Section 7.7 of this paper for a definition of the term sufficiently
local.
17
49
10.1. Embedding classical (memoryless) computation in quantum

mechanics.
One objective in this section is to represent18 classical computing computing devices as unitary transformations. Since unitary transformations
are invertible, i.e., reversible, it follows that the only classical computing
devices that can be represented as such transformations must of necessity
be reversible devices. Hence, the keen interest in reversible computation.
For a more in depth study of reversible computation, please refer to the
work of Bennett and others.
10.2. Classical reversible computation without memory.
Input
xn1
xn1
..
.
..
.
x1
x0
..
.
CRCDn
yn1
yn1
..
.
y1
y0
Output
Each classical n-input/n-output (binary memoryless) reversible computing device (CRCDn) can be identified with a bijection
: {0, 1}n {0, 1}n
on the set {0, 1}n of all binary n-tuples. Thus, we can in turn identify each
CRCDn with an element of the permutation group S2n on the 2n symbols
{ h
a| |
a {0, 1}n } .
Let
Bn = B hx0, x1, . . . , xn1 i
denote the free Boolean ring on the symbols x0 , x1, . . . , xn1 . Then
the binary n-tuples

a {0, 1}n are in one-to-one correspondence with the
minterms of Bn , i.e.,
a x a =
where
n1
Y
xj j ,
j=0
0
xj = xj
x1j = xj
Since there is a one-to-one correspondence between the automorphisms of

Bn and the permutations on the set of minterms, it follows that CRCDn s
18
Double meaning is intended.
50
can also be identified with the automorphism group Aut (Bn ) of the free
Boolean ring Bn .
Moreover, since the set of binary n-tuples {0, 1}n is in one-to-one correspondence with the set of integers {0, 1, 2, . . . , 2n 1} via the radix 2
representation of integers, i.e.,
(bn1 , bn2 , . . . , b1, b0)
n1
X
bj 2j ,
j=0
we can, and frequently do, identify binary n-tuples with integers.

For example, consider the Controlled-NOT gate, called CNOT , which
is defined by the following wiring diagram:
c
CNOT =
b+c
b
where and denote respectively a control bit and a target bit, and
where a + b denotes the exclusive or of bits a and b. This corresponds to
the permutation = (26)(37), i.e.,
|0i = |000i 7 |000i = |0i
|1i = |001i 7 |001i = |1i
|2i
= |010i 7 |110i = |6i
|3i = |011i 7 |111i = |7i

,
|4i
=
|100i
|100i
=
|4i
|5i = |101i 7 |101i = |5i
|6i
= |110i 7 |010i = |2i
|7i = |111i 7 |011i = |3i

where we have used the following indexing conventions:
First=Right=Bottom
Last=Left=Top
As another example, consider the Toffoli gate , which is defined by the

following wiring diagram:
Toffoli =
c c + ab
|
b
b
|
a
a
51
where ab denotes the logical and of a and b. As before, + denotes

exclusive or. This gate corresponds to the permutation = (67).
In summary, we have:
{ CRCDn } = S2n = Aut (Bn )
10.3. Embedding classical irreversible computation within classical

reversible computation.
A classical 1-input/n-output (binary memoryless) irreversible computing
device can be thought of as a Boolean function f = f (xn2 , . . . , x1, x0) in
Bn1 = B hx0 , x1, . . . , xn2 i. Such irreversible computing devices can be
transformed into reversible computing devices via the monomorphism
: Bn1 Aut(Bn ),
where (f ) is the automorphism in Aut(Bn ) defined by

(xn1 , xn2 , . . . , x1, x0) 7 (xn1 f, xn2 , . . . , x1, x0) ,
and where denotes exclusive or. Thus, the image of each Boolean
function f is a product of disjoint transpositions in S2n .
As an additive group (ignoring ring structure), Bn1 is the abelian group
L2(n1) 1
Z2 , where Z2 denotes the cyclic group of order two.
j=0
Classical Binary Memoryless Computation is summarized in the table

below:
Summary
Classical Binary Memoryless Computation
Bn1 =
L2(n1) 1
j=0
Z2 S2n = Aut(Bn )
10.4. The unitary representation of reversible computing devices.

It is now a straight forward task to represent CRCDn s as unitary transformations. We simply use the standard unitary representation
: S2n U(2n ; C)
of the symmetric group S2n into the group of 2n 2n unitary matrices
U(2n ; C). This is the representation defined by
7 (k,k )2n 2n ,
52
where k` denotes the Kronecker delta, i.e.,
1 if k = `
k` =
0 otherwise
We think of such unitary transformations as quantum computing devices.

For example, consider the controlled-NOT gate CNOT0 = (45)(67) S8
given by the wiring diagram

c
|
b
b
|
a a + c
c
CNOT0 =
This corresponds to the unitary transformation
1 0 0
0 1 0
0 0 1
0 0 0
0
UCNOT0 = (CNOT ) =
0 0 0
0 0 0
0 0 0
0 0 0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
Moreover, consider the Toffoli gate Toffoli0 = (57) S8 given by the

wiring diagram

c
|
b b + ac
|
a
a
c
Toffoli0 =
This corresponds to the unitary transformation
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
0
0
UToffoli = (Toffoli ) =
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
53
Abuse of Notation and a Caveat: Whenever it is clear from context,

we will use the name of a CRCDn to also refer to the unitary transformation corresponding to the CRCDn . For example, we will denote
(CN OT ) and (T of f oli) simply by CN OT and T of f oli. Moreover
we will also use the wiring diagram of a CRCDn to refer to the unitary
transformation corresponding to the CRCDn . For quantum computation beginners, this can lead to some confusion. Be careful!
10.5. Some other simple quantum computing devices.
After CRCDn s are embedded as quantum computing devices, they are
no longer classical computing devices. After the embedding, they suddenly
have acquired much more computing power. Their inputs and outputs can
be a superposition of many states. They can entangle their outputs. It
is misleading to think of their input qubits as separate, for they could be
entangled.
As an illustration of this fact, please note that the
device CNOT00 given by the wiring diagram
1
b a + b
0
|
=
CNOT00 =
0
a
a
0
quantum computing
0
1
0
0
0
0
0
1
0
0
1
0
is far from classical. It is more than a permutation. It is a linear operator

that respects quantum superposition.
For example, CNOT00 can take two non-entangled qubits as input, and
then produce two entangled qubits as output. This is something no classical
computing device can do. For example,
|0i |1i
1
1
|0i = (|00i |10i) 7 (|00i |11i)

2
2
2
For completeness, we list two other quantum computing devices that are
embeddings of CRCDn s, NOT and SWAP:

0 1
= 1
NOT = a NOT a + 1 =
1 0
and
a
|
|
|
a b
b
SWAP =
1
0
=
0
0
0
0
1
0
0
1
0
0
0
0
0
1
54
10.6. Quantum computing devices that are not embeddings.

We now consider quantum computing devices that are not embeddings of
CRCDn s.
The Hadamard gate H is defined as:

H=
1
=
2
1
1
1 1
Another quantum gate is the square root of NOT, i.e., NOT, which
is given by

1i
1+i
i 1
1 i
NOT = NOT =
=
.
1 i
i
1
2
2
SWAP =
SWAP
SWAP which is defined as:
1 0
0 0
0 1+i 1i 0
2
2
=
0 1i 1+i 0 .
2
2
0 0
0 1
There is also the square root of swap
Three frequently used unary quantum gates are the rotations:

cos i sin
i
= ei1
e 1 = i sin cos
i2
i3
cos sin
sin cos
ei
0
0 ei
= ei2
= ei3
10.7. The implicit frame of a wiring diagram.

Wiring diagrams have the advantage of being a simple means of describing
some rather complicated unitary transformations. However, they do have
their drawbacks, and they can, if we are not careful, be even misleading.
One problem with wiring diagrams is that they are not frame (i.e., basis)
independent descriptions of unitary transformations. Each wiring diagram
describes a unitary transformation using an implicitly understood basis.
55
For example, consider CNOT00 given by the wiring diagram:

00
CNOT =
a + b
|
a
a
b
The above wiring diagram defines CNOT00 in terms of the implicitly understood basis

0
1
.
, |1i =
|0i =
1
0
This wiring diagram suggests that qubit #1 controls qubit #0, and that
qubit #1 is not effected by qubit #0. But this is far from the truth. For,
CNOT00 transforms
|0i + |1i |0i |1i

2
2
into
|0i |1i |0i |1i

,
2
2
where we have used our indexing conventions
First=Right=Bottom
.
Last=Left=Top
In fact, in the basis

0 |0i + |1i 0 |0i |1i
0 =

, 1 =
2
2
the wiring diagram of the same unitary transformation CNOT00 is:
b a + b
|
a
a
The roles of the target and control qubits appeared to have switched!
11. The No-Cloning Theorem
In this section, we prove the no-cloning theorem of Wootters and Zurek

[83]. The theorem states that there can be no device that produces exact
replicas or copies of a quantum state. (See also [83] for an elegant proof
using the creation operators of quantum electrodynamics.)
The proof is an amazingly simple application of the linearity of quantum mechanics. The key idea is that copying is an inherently quadratic
56
transformation, while the unitary transformations of quantum mechanics

are inherently linear. Ergo, copying can not be a unitary transformation.
But what do we mean by a quantum replicator?
Definition 11. Let H be a Hilbert space. Then a quantum replicator
consists of an auxiliary Hilbert space HA , a fixed state |0i HA (called the
initial state of replicator), and a unitary transformation
U : HA H H HA H H
such that, for some fixed state |blanki H,
U |0i |ai |blanki = |a i |ai |ai ,
for all states |ai H, where |ai HA (called the replicator state after
replication of |ai) depends on |ai.
Since a quantum state is determined by a ket up to a multiplicative nonzero complex number, we can without loss of generality assume that |0i,
|ai, |blanki are all of unit length. From unitarity, it follows that |ai is also
of unit length.
Let |ai, |bi be two kets of unit length in H such that
0<|ha|bi|<1.
Then
U |0i |ai |blanki = |ai |ai |ai
Hence,
U |0i |bi |blanki =
|b i |bi |bi
hblank| ha| h0| U U |0 i |bi |blanki = hblank| ha| h 0 | 0 i |bi |blanki

=ha|bi
On the other hand,

hblank| ha| h0| U U |0i |bi |blanki = ha| ha| h a | b i |bi |bi
= h a | b i 2 h a | b i
Thus,
And so,
h a | b i 2 h a | b i = h a | b i .
h a | b i h a | b i = 1 .
57
But this equation can not be satisfied since

|h a | b i| < 1
and
|h a | b i| k |ai k k |b i k = 1
Hence, a quantum replicator cannot exist.
12. Quantum teleportation
We now give a brief description of quantum teleportation, a means possibly to be used by future quantum computers to bus qubits from one location
to another.
As stated earlier, qubits can not be copied as a result of the no-cloning
theorem. (Please refer to the previous section.) However, they can be
teleported, as has been demonstrated in laboratory settings. Such a mechanism could be used to bus qubits from one computer location to another.
It could be used to create devices called quantum repeaters.
But what do we mean by teleportation?

Teleportation is the transferring of an object from one location to another by a process that:
1) Firstly dissociates (i.e., destroys) the object to obtain information.
The object to be teleported is first scanned to extract sufficient information to reassemble the original object.
2) Secondly transmits the acquired information from one location to another.
3) Lastly reconstructs the object at the new location from the received
information. An exact replicas re-assembled at the destination out
of locally available materials.
Two key effects of teleportation should be noted:
1) The original object is destroyed during the process of teleportation.
Hence, the no-cloning theorem is not violated.
2) An exact replica of the original object is created at the intended destination.
58
Scotty of the Starship Enterprise was gracious enough to loan me the

following teleportation manual. So I am passing it on to you.
Quantum Teleportation Manual

Step. 1 .(Location A): Preparation: At location A, construct an EPR
pair of qubits (qubits #2 and #3) in H2 H3 .
|00i 7
Unitary
Matrix
H2 H 3
|01i|10i
H2 H 3
Step 2. Transport: Physically transport entangled qubit #3 from location A to location B.

Step 3. : The qubit to be teleported, i.e., qubit #1, is delivered to location A in an unknown state
a |0i + b |1i
As a result of Steps 1 - 3, we have:
Locations A and B share an EPR pair, i.e.
The qubit which is to be teleported, i.e., qubit #1, is at Location
A
Qubit #2 is at Location A
Qubit #3 is at Location B
Qubits #2 & #3 are entangled
The current state |i of all three qubits is:

|01i |10i
|i = (a |0i + b |1i)
H1 H 2 H3
2
To better understand what is about to happen, we re-express the state
|i of the three qubits in terms of the following basis (called the Bell basis)
of H1 H2 :
|A i = |10i|01i
|10i+|01i
|B i =
2
|C i =
|D i =
|00i|11i
|00i+|11i
59
The result is:

1
2[
|i =
|A i (a |0i b |1i)
+ |B i (a |0i + b |1i)
+ |C i (a |1i + b |0i)
+ |D i (a |1i b |0i) ]
where, as you might have noticed, we have written the expression in a suggestive way.
Remark 11. Please note that since the completion of Step 3, we have done
nothing physical. We have simply performed some algebraic manipulation
of the expression representing the state |i of the three qubits.
Let U : H1 H2 H1 H2 be the unitary transformation defined by
|A i 7 |00i
|B i 7 |01i
|C i
7 |10i
|D i
7 |11i
Step 4. (Location A): 19 Apply the local unitary transformation U

I : H1 H2 H3 H1 H2 H3 to the three qubits (actually more
precisely, to qubits #1 and #2). Thus, under U I the state |i of
all three qubits becomes
1
2[
|0i =
|00i (a |0i b |1i)

+ |01i (a |0i + b |1i)
+ |10i (a |1i + b |0i)
+ |11i (a |1i b |0i) ]
Step 5. (Location A): Measure qubits #1 and #2 to obtain two bits

of classical information. The result of this measurement will be one
of the bit pairs {00, 01, 10, 11}.
Step 6.: Send from location A to location B (via a classical communication channel) the two classical bits obtained in Step 6.
19
Actually, there is no need to apply the unitary transformation U . We could have
instead made a complete Bell state measurement, i.e., a measurement with respect to the
compatible observables |Ai hA |, |B i hB |, |C i hC |, |D i hD |. We have added
the additional step 4 to make quantum teleportation easier to understand for quantum
computation beginners.
Please note that a complete Bell state measurement has, of this writing, yet to be
achieved in a laboratoy setting.
60
As an intermediate summary, we have:

1) Qubit #1 has been disassembled, and
2) The information obtained during disassembly (two classical bits) has
been sent to location B.
Step 7. (Location B): The two bits (i, j) received from location A are
used to select from the following table a unitary transformation U (i,j)
of H3 , (i.e., a local unitary transformation I4 U (i,j) on H1 H2 H3 )
Rec. Bits
00
01
10
11
U (i,j)

1
U (00) =
0

1
U (01) =
0

0
U (10) =
1

0
U (11) =
1
0
1

0
1

1
0

1
0
Future effect on qubit #3

a |0i b |1i 7 a |0i + b |1i
a |0i + b |1i 7 a |0i + b |1i
a |1i + b |0i 7 a |0i + b |1i
a |1i b |0i 7 a |0i + b |1i
Step 8. (Location B): The unitary transformation U (i,j) selected in

Step 7 is applied to qubit #3.
As a result, qubit #3 is at location B and has the original state of qubit
#1 when qubit #1 was first delivered to location A, i.e., the state
a |0i + b |1i
It is indeed amazing that no one knows the state of the quantum teleported qubit except possibly the individual that prepared the qubit. Knowledge of the actual state of the qubit is not required for teleportaton. If its
state is unknown before the teleportation, it remains unknown after the
teleportation. All that we know is that the states before and after the
teleportation are the same.
13. Shors algorithm

The following description of Shors algorithm is based on [27], [47], [50],
[53], and [77] .
61
13.1. Preamble to Shors algorithm.

There are cryptographic systems (such as RSA20 ) that are extensively
used today (e.g., in the banking industry) which are based on the following
questionable assumption, i.e., conjecture:
Conjecture(Assumption). Integer factoring is computationally much
harder than integer multiplication. In other words, while there are obviously
many polynomial time algorithms for integer multiplication, there are no
polynomial time algorithms for integer factoring. I.e., integer factoring
computationally requires super-polynomial time.
This assumption is based on the fact that, in spite of the intensive efforts
over many centuries of the best minds to find a polynomial time factoring
algorithm, no one has succeeded so far. As of this writing, the most asymptotically efficient classical algorithm isthe number
theoretic sievei[56], [57],
h
1/3
which factors an integer N in time O exp (lg N ) (lg lg N )2/3 . Thus,
this is a super-polynomial time algorithm in the number O (lg N ) of digits
in N .
However, ... Peter Shor suddenly changed the rules of the game.
Hidden in the above conjecture is the unstated, but implicitly understood,
assumption that all algorithms run on computers based on the principles of
classical mechanics, i.e., on classical computers. But what if a computer
could be built that is based not only on classical mechanics, but on quantum
mechanics as well? I.e., what if we could build a quantum computer?
Shor, starting from the works of Benioff, Bennett, Deutsch , Feynman,
Simon, and others, created an algorithm to be run on a quantum computer, i.e., a quantum algorithm, that
factors integers in polynomial
time!
2
Shors algorithm takes asymptotically O (lg N ) (lg lg N ) (lg lg lg N ) steps
on a quantum computer, which is polynomial time in the number of digits
O (lg N ) of N .
20
RSA is a public key cryptographic system invented by Rivest, Shamir, Adleman.

Hence the name. For more information, please refer to [79].
62
13.2. Number theoretic preliminaries.

Since the time of Euclid, it has been known that every positive integer N
can be uniquely (up to order) factored into the product of primes. Moreover,
it is a computationally easy (polynomial time) task to determine whether or
not N is a prime or composite number. For the primality testing algorithm
of Miller-Rabin[66] makes such a determination at the cost of O (s lg N )
arithmetic operations [O s lg3 N bit operations] with probability of error
P robError 2s .
However, once an odd positive integer N is known to be composite, it does
not appear to be an easy (polynomial time) task on a classical computer to
determine its prime factors. As mentioned earlier, so far the most asymptotically efficient classical algorithm known isthe number
theoretic sievei[56],
h
1/3
[57], which factors an integer N in time O exp (lg N ) (lg lg N )2/3 .
Prime Factorization Problem. Given a composite odd positive integer

N , find its prime factors.
It is well known[66] that factoring N can be reduced to the task of choosing
at random an integer m relatively prime to N , and then determining its
modulo N multiplicative order P , i.e., to finding the smallest positive integer
P such that
mP = 1 mod N .
It was precisely this approach to factoring that enabled Shor to construct
his factoring algorithm.
13.3. Overview of Shors algorithm.

But what is Shors quantum factoring algorithm?
Let N = {0, 1, 2, 3, . . . } denote the set of natural numbers.
Shors algorithm provides a solution to the above problem. His algorithm

consists of the five steps (steps 1 through 5), with only STEP 2 requiring
the use of a quantum computer. The remaining four other steps of the
algorithm are to be performed on a classical computer.
We begin by briefly describing all five steps. After that, we will then
focus in on the quantum part of the algorithm, i.e., STEP 2.
63
Step 1. Choose a random positive even integer m. Use the polynomial time
Euclidean algorithm21 to compute the greatest common divisor gcd (m, N )
of m and N . If the greatest common divisor gcd (m, N ) 6= 1, then we
have found a non-trivial factor of N , and we are done. If, on the other
hand, gcd (m, N ) = 1, then proceed to STEP 2.
STEP 2. Use a quantum computer to determine the unknown period P of
the function
f
N
N
N
a 7 ma mod N
Step 3. If P is an odd integer, then goto Step 1. [The probability of P being

odd is ( 21 )k , where k is the number of distinct prime factors of N .] If
P is even, then proceed to Step 4.
Step 4. Since P is even,

mP/2 1 mP/2 + 1 = mP 1 = 0 mod N .
If mP/2 + 1 = 0 mod N , then goto Step 1. If mP/2 + 1 6= 0 mod N ,

then proceed to Step 5. It can be shown that the probability that
mP/2 + 1 = 0 mod N is less than ( 21 )k1 , where k denotes the number
of distinct prime factors of N .

Step 5. Use the Euclidean algorithm to compute d = gcd mP/2 1, N . Since
mP/2 +1 6= 0 mod N , it can easily be shown that d is a non-trivial factor

of N . Exit with the answer d.
Thus, the task of factoring an odd positive integer N reduces to the

following problem:
Problem. Given a periodic function
find the period P of f .
21
f : N N ,
The Euclidean algorithm is O lg2 N . For a description of the Euclidean algorithm,

see for example [20] or [19].
64
13.4. Preparations for the quantum part of Shors algorithm.

Choose a power of 2
Q = 2L
such that
N 2 Q = 2L < 2N 2 ,
and consider f restricted to the set
SQ = {0, 1, . . . , Q 1}
which we also denote by f , i.e.,
f : SQ SQ .
In preparation for a discussion of STEP 2 of Shors algorithm, we construct two L-qubit quantum registers, Register1 and Register2 to hold
respectively the arguments and the values of the function f , i.e.,
|Reg1i |Reg2i = |ai |f (a)i = |ai |bi = |a0a1 aL1 i |b0 b1 bL1 i
In doing so, we have adopted the following convention for representing
integers in these registers:
Notation Convention. In a quantum computer, we represent an integer
a with radix 2 representation
a=
L1
X
aj 2j ,
j=0
as a quantum register consisting of the 2n qubits

|ai = |a0a1 aL1 i =
L1
O
j=0
|aj i
For example, the integer 23 is represented in our quantum computer as n

qubits in the state:
|23i = |10111000 0i
Before continuing, we remind the reader of the classical definition of the

Q-point Fourier transform.
65
Definition 12. Let be a primitive Q-th root of unity, e.g., = e2i/Q.

Then the Q-point Fourier transform is the map
F
M ap(SQ, C) M ap(SQ , C)
h
i
[f : SQ C] 7 fb : SQ C
where
1 X
f (x) xy
fb(y) =
Q
xSQ
We implement the Fourier transform F as a unitary transformation, which

in the standard basis
|0i , |1i , . . . , |Q 1i
is given by the Q Q unitary matrix

1
F = ( xy ) .
Q

This unitary transformation can be factored into the product of O lg2 Q =

O lg2 N sufficiently local unitary transformations. (See [77], [47].)
13.5. The quantum part of Shors algorithm.
The quantum part of Shors algorithm, i.e., STEP 2, is the following:
STEP 2.0 Initialize registers 1 and 2, i.e.,
STEP 2.1
|0i = |Reg1i |Reg2i = |0i |0i = |00 0i |0 0i

22 Apply
the Q-point Fourier transform F to Register1.
Q1
Q1
1 X 0x
1 X
|0i = |0i |0i 7 |1i =
|xi |0i =
|xi |0i
Q x=0
Q x=0
FI
Remark 12. Hence, Register1 now holds all the integers

0, 1, 2, . . . , Q 1
in superposition.
22
In this step we could have instead applied the Hadamard transform to Register1
with the same result, but at the computational cost of O (lg N ) sufficiently local unitary
transformations.
66
STEP 2.2 Let Uf be the unitary transformation that takes |xi |0i to |xi |f (x)i.
Apply the linear transformation Uf to the two registers. The result
is:
Q1
Q1
Uf
1 X
1 X
|xi |0i 7 |2i =
|xi |f (x)i
|1i =
Q x=0
Q x=0
Remark 13. The state of the two registers is now more than a superposition
of states. In this step, we have quantum entangled the two registers.
STEP 2.3. Apply the Q-point Fourier transform F to Reg1. The resulting state
is:
|2i =
1
Q
Q1
X
x=0
FI
|xi |f (x)i 7 |3 i =
1
Q
1
Q
Q1
X
X Q1
x=0 y=0
Q1
X
y=0
xy |yi |f (x)i
|(y)i
k|(y)ik |yi k|(y)ik ,
where
|(y)i =
Q1
X
x=0
xy |f (x)i .
STEP 2.4. Measure Reg1, i.e., perform a measurement with respect to the orthogonal projections
|0i h0| I, |1i h1| I, |2i h2| I, . . . , |Q 1i hQ 1| I ,
where I denotes the identity operator on the Hilbert space of the second
register Reg2.
As a result of this measurement, we have, with probability
P rob (y0 ) =
k|(y0 )ik2
,
Q2
moved to the state

|y0 i
and measured the value
|(y0 )i
k|(y0 )ik
y0 {0, 1, 2, . . . , Q 1} .
67
If after this computation, we ignore the two registers Reg1 and Reg2, we
see that what we have created is nothing more than a classical probability
distribution S on the sample space
{0, 1, 2, . . . , Q 1} .
In other words, the sole purpose of executing STEPS 2.1 to 2.4 is to create
a classical finite memoryless stochastic source S which outputs a symbol
y0 {0, 1, 2, . . . , Q 1} with the probability
P rob(y0) =
k|(y0 )ik2
.
Q2
(For more details, please refer to section 8.1 of this paper.)
As we shall see, the objective of the remander of Shors algorithm is to

glean information about the period P of f from the just created stochastic
source S. The stochastic source was created exactly for that reason.
13.6. Peter Shors stochastic source S.
Before continuing to the final part of Shors algorithm, we need to analyze

the probability distribution P rob (y) a little more carefully.
Proposition 1. Let q and r be the unique non-negative integers such that

Q = P q + r , where 0 r < P ; and let Q0 = P q. Then
P rob (y) =
r sin2
P y
Q0
+1
P
+(P r) sin2
Q2 sin2
r(Q0 +P ) 2 +(P r)Q20
Q2 P 2
P y
Q
P y Q0
P
Q
if P y 6= 0 mod Q
if P y = 0 mod Q
68
Proof. We begin by deriving a more usable expression for |(y)i.

|(y)i =
Q1
X
P
1
X
xy
x=0
|f (x)i =
Q0
1
P
x0 =0 x1 =0
P
1
X
x0 =0
r1
X
x0 =0
x0 y
QX
0 1
xy
x=0
|f (x)i +
Q1
X
x=Q0
(P x1 +x0 )y |f (P x1 + x0 )i +
xy |f (x)i
r1
X
Q0
P
x0 =0
Q
0 1
r1
P
X
Py
X P yx1
x0 y
|f (x0)i +
x1 =0
Q0
P
Q0
P
x0 =0
X P yx1
x0 y
|f (x0)i +
x1 =0
i
+x0 y
P
1
X
x0 =r
x0 y
|f (P x1 + x0 )i
|f (x0)i
Q
0 1
P
X P yx1
|f (x0)i
x1 =0
where we have used the fact that f is periodic of period P .
Since f is one-to-one when restricted to its period 0, 1, 2, . . . , P 1, all

the kets
|f (0)i , |f (1)i , |f (2)i , . . . , |f (P 1)i ,
are mutually orthogonal. Hence,

2
2
Q0 1
Q0

PX
X

P

h(y) | (y)i = r
P yx1 .
P yx1 + (P r)

x1 =0
x1 =0

If P y = 0 mod Q, then since is a Q-th root of unity, we have

2
2
Q0
Q0
h(y) | (y)i = r
+ 1 + (P r)
.
P
P
On the other hand, if P y 6= 0 mod Q, then we can sum the geometric

series to obtain

2

2
P y QP0 +1

P y QP0

1
+ (P r))

h(y) | (y)i =

P
y
P
y
1

1

2
2
Q0
Q0
2i

2i
e Q P y P 1
e Q P y P +1 1

=
2i
2i
+ (P r))

P y
P y
Q
Q

e
1
e
1
69
where we have used the fact that is the primitive Q-th root of unity given
by
= e2i/Q .
The remaining part of the proposition is a consequence of the trigonometric identity

2

i
2
.
e 1 = 4 sin
2
As a corollary, we have
Corollary 1. If P is an exact divisor of Q, then
0 if P y 6= 0 mod Q
P rob (y) =
1
if P y = 0 mod Q
P
13.7. A momentary digression: Continued fractions.
We digress for a moment to review the theory of continued fractions. (For
a more in-depth explanation of the theory of continued fractions, please refer
to [42] and [58].)
Every positive rational number can be written as an expression in the
form
1
,
= a0 +
1
a1 +
1
a2 +
1
a3 +
1
+
aN
where a0 is a non-negative integer, and where a1 , . . . , aN are positive integers. Such an expression is called a (finite, simple) continued fraction ,
and is uniquely determined by provided we impose the condition aN > 1.
For typographical simplicity, we denote the above continued fraction by
[a0 , a1, . . . , aN ] .
The continued fraction expansion of can be computed with the following
recurrence relation, which always terminates if is rational:
a0 = bc
an+1 = b1/n c
, and if n 6= 0, then

1
0 = a0
n+1 = n an+1
70
The n-th convergent (0 n N ) of the above continued fraction is

defined as the rational number n given by
n = [a0 , a1, . . . , an ] .
Each convergent n can be written in the for, n = pqnn , where pn and qn
are relatively prime integers ( gcd (pn , qn ) = 1). The integers pn and qn are
determined by the recurrence relation
p0 = a0 , p1 = a1 a0 + 1, pn = an pn1 + pn2 ,
q0 = 1,
q 1 = a1 ,
qn = an qn1 + qn2 .
13.8. Preparation for the final part of Shors algorithm.

Definition 13. 23 For each integer a, let {a}Q denote the residue of a
modulo Q of smallest magnitude. In other words, {a}Q is the unique
integer such that
a = {a}Q mod Q
.
Q/2 < {a}Q Q/2

Proposition 2. Let y be an integer lying in SQ. Then

4
1
1 2
if 0 < {P y}Q
2 P 1 N
P rob (y)

1
1 2
if {P y}Q = 0
P 1 N
Proof. We begin by noting that
{P y}

Q
P
2 1
Q Q P0 + 1 Q
1
N
1
N
Q0 +P

1+
P
Q
where we have made use of the inequalities
It immediately follows that

{P y}Q Q0

<
.

Q
P 2
j
23
{a}Q = a Q round
a
Q
=aQ
a
Q
1
2
N 2 Q < 2N 2 and 0 < P N .
P
2
1
N
1
N
1
N
Q+P
1+
N
N2
<
71
As a result, we can legitimately use the inequality

4 2
sin2 2 , for || <

2
2
to simplify the expression for P rob (y).
Thus,
{P y}Q
r sin2
P rob (y) =
+(P r) sin2
{P y}Q
Q0
+1
P
+(P r)
4
2
4
2
1
P
Q0
P
Q2
Qr
Q
4
2
1
P

1
4
2
1
P
r
Q
2
{P y}Q Q
P0
Q
{P y}Q
Q
Q2
{P y}Q Q
P0
Q
P y
Q
Q2 sin2
Q0
+1
P
2

1 2
N
The remaining case, {P y}Q = 0 is left to the reader.

Lemma 1. Let

P

Y = y SQ | {P y}Q
2
and
SP = {d SQ | 0 d < P } .
Then the map
Y
y
SP

P
7 d = d(y) = round Q
y
is a bijection with inverse
y = y(d) = round
Q
d
P
Hence, Y and SP are in one-to-one correspondence. Moreover,

{P y}Q = P y Q d(y) .
Remark 14. Moreover, the following two sets of rationals are in one-to-one
correspondence

d
y
| y Y
|0d<P
Q
P
72
As a result of the measurement performed in STEP 2.4, we have in our

possession an integer y Y . We now show how y can be use to determine
the unknown period P .
We now need the following theorem24 from the theory of continued fractions:
Theorem 2. Let be a real number, and let a and b be integers with b > 0.
If

a
1

2 ,
b
2b
then the rational number a/b is a convergent of the continued fraction expansion of .
As a corollary, we have:

Corollary 2. If {P y}Q P2 , then the rational number
y
.
of the continued fraction expansion of Q
d(y)
P
is a convergent
Proof. Since
P y Qd(y) = {P y}Q ,
we know that
|P y Qd(y)|
which can be rewritten as
P
,
2

y
d(y) 1 .
Q
P 2Q
But, since Q N 2, it follows that

y

d(y) 1 .
Q
P 2N 2
Finally, since P N (and hence
applied. Thus,
y
=Q
.
d(y)
P
1
2N
1
),
2P 2
the above theorem can be
is a convergent of the continued fraction expansion of
d(y)
Since P is a convergent of the continued fraction expansion of

follows that, for some n,
d(y)
pn
,
=
P
qn
24
See [42, Theorem 184, Section 10.15].
y
Q,
it
73
where pn and qn are relatively prime positive integers given by a recurrence

relation found in the previous subsection. So it would seem that we have
found a way of deducing the period P from the output y of STEP 2.4, and
so we are done.
Not quite!
We can determine P from the measured y produced by STEP 2.4, only if
pn = d(y)
,
qn = P
which is true only when d(y) and P are relatively prime.
So what is the probability that the y Y produced by STEP 2.4 satisfies

the additional condition that
gcd (P, d(y)) = 1 ?
Proposition 3. The probability that the random y produced by STEP 2.4 is

such that d(y) and P are relatively prime is bounded below by the following
expression

4 (P )
1 2
P rob {y Y | gcd(d(y), P ) = 1} 2
1
,
P
N
where (P ) denotes Eulers totient function, i.e., (P ) is the number of
positive integers less than P which are relatively prime to P .
The following theorem can be found in [42, Theorem 328, Section 18.4]:
Theorem 3.
lim inf
(N )
= e ,
N/ ln ln N
where denotes Eulers constant = 0.57721566490153286061 . . . , and

where e = 0.5614594836 . . . .
As a corollary, we have:
Corollary 3.
P rob {y Y | gcd(d(y), P ) = 1}

e (P )
1 2
4
,
2 ln 2
lg lg N
N
74
where (P ) is a monotone decreasing sequence converging to zero. In terms

of asymptotic notation,

1
.
P rob {y Y | gcd(d(y), P ) = 1} =
lg lg N
Thus , if STEP 2.4 is repeated O(lg lg N ) times, then the probability of
success is (1).
Proof. From the above theorem, we know that
(P )
e (P ) .
P/ ln ln P
where (P ) is a monotone decreasing sequence of positive reals converging
to zero. Thus,
e (P )
e (P )
e (P )
e (P )
1
(P )
P
ln ln P
ln ln N
ln ln 2 + ln lg N
ln 2
lg lg N
Remark 15. ( lg lg1 N ) denotes an asymptotic lower bound. Readers not

familiar with the big-oh O() and big-omega () notation should refer to
[19, Chapter 2] or [11, Chapter 2].
Remark 16. For the curious reader, lower bounds LB(P ) of e (P )
for 3 P 841 are given in the following table:
P
3
4
5
7
13
31
61
211
421
631
841
LB(P )
0.062
0.163
0.194
0.303
0.326
0.375
0.383
0.411
0.425
0.435
0.468
Thus, if one wants a reasonable bound on the P rob {y Y | gcd(d(y), P ) = 1}

before continuing with Shors algorithm, it would pay to first use a classical
algorithm to verify that the period P of the randomly chosen integer m is
not too small.
75
13.9. The final part of Shors algorithm.

We are now prepared to give the last step in Shors algorithm. This step
can be performed on a classical computer.
Step 2.5 Compute the period P from the integer y produced by STEP 2.4.
Loop for each n from n = 1 Until n = 0.
Use the recurrence relations given in subsection 13.7, to comy

pute the pn and qn of the n-th convergent pqnn of Q
.
Test to see if qn = P by computing25

Y i qn,i
mqn =
m2
mod N ,
i
where qn = i qn,i 2i is the binary expansion of qn .

If mqn = 1 mod N , then exit with the answer P = qn , and
proceed to Step 3. If not, then continue the loop.
End of Loop
If you happen to reach this point, you are a very unlucky quantum
computer scientist. You must start over by returning to STEP
2.0. But dont give up hope! The probability that the integer y
produced by STEP 2.4 will lead to a successful completion of Step
2.5 is bounded below by

4
e (P )
1 2
0.232
1 2
1
>
1
,
2 ln 2
lg lg N
N
lg lg N
N
provided the period P is greater than 3.
constant.]
25
[ denotes Eulers
The indicated algorithm for computing mqn mod N requires O(lg qn ) arithmetic
operations.
76
13.10. An example of Shors algorithm.
Let us now show how N = 91 (= 7 13) can be factored using Shors

algorithm.
We choose Q = 214 = 16384 so that N 2 Q < 2N 2.
Step 1 Choose a random positive integer m, say m = 3. Since gcd(91, 3) = 1,
we proceed to STEP 2 to find the period of the function f given by
f (a) = 3a mod 91
Remark 17. Unknown to us, f has period P = 6. For,
a
f (a)
0 1 2
6 7
1 3 9 27 81 61 1 3
Unknown period P = 6
STEP 2.0 Initialize registers 1 and 2. Thus, the state of the two registers becomes:
|0i = |0i |0i
STEP 2.1 Apply the Q-point Fourier transform F to register #1, where
16383
X
1
F |ki =
kj |xi ,
16384 x=0
2i
and where is a primitive Q-th root of unity, e.g., = e 16384 . Thus

the state of the two registers becomes:
16383
X
1
|1i =
|xi |0i
16384 x=0
77
STEP 2.2 Apply the unitary transformation Uf to registers #1 and #2, where
Uf |xi |ì = |xi | f (x) ` mod 91i .
(Please note that Uf2 = I.) Thus, the state of the two registers becomes:
|2 i =
1
16384
1
(
16384
P16383
x=0
|xi |3x mod 91i
| 0i |1i + | 1i |3i + | 2i |9i + | 3i |27i + | 4i |81i + | 5i |61i
+ | 6i |1i + | 7i |3i + | 8i |9i + | 9i |27i + |10i |81i + |11i |61i

+ |12i |1i + |13i |3i + |14i |9i + |15i |27i + |16i |81i + |17i |61i
+ ...
+ |16380i |1i + |16381i |3i + |16382i |9i + |16383i |27i
Remark 18. The state of the two registers is now more than a superposition
of states. We have in the above step quantum entangled the two registers.
STEP 2.3 Apply the Q-point F again to register #1. Thus, the state of the
system becomes:
|3 i =
where
1
16384
1
16384
1
16384
P16383
x=0
P16383
x=0
P16383
x=0
| (y)i =
1
16384
|yi
P16383
P16383
x=0
y=0
xy |yi |3x mod 91i
xy |3x mod 91i
|yi | (y)i ,
16383
X
x=0
xy |3x mod 91i
78
Thus,
|1i + y |3i + 2y |9i + 3y |27i + 4y |81i + 5y |61i
| (y)i =
+ 6y |1i + 7y |3i + 8y |9i + 9y |27i + 10y |81i + 11y |61i

+ 12y |1i + 13y |3i + 14y |9i + 15y |27i + 16y |81i + 17y |61i
+ ...
+ 16380y |1i + 16381y |3i + 16382y |9i + 16383y |27i
STEP 2.4 Measure Reg1. The result of our measurement just happens to turn
out to be
y = 13453
Unknown to us, the probability of obtaining this particular y is:
0.3189335551 106 .
Moreover, unknown to us, were lucky!

prime to P , i.e.,
d = d(y) = round(
The corresponding d is relatively

P
y) = 5
Q
However, we do know that the probability of d(y) being relatively prime

to P is greater than

0.232
1 2
8.4% (provided P > 3),
1
lg lg N
N
and we also know that
d(y)
P
is a convergent of the continued fraction expansion of
y
13453
=
=
Q
16384
So with a reasonable amount of confidence, we proceed to Step 2.5.
Step 2.5 Using the recurrence relations found in subsection 13.7 of this paper,
we successively compute (beginning with n = 0) the an s and qn s for
the continued fraction expansion of
13453
y
=
.
=
Q
16384
79
For each non-trivial n in succession, we check to see if

3qn = 1 mod 91.
If this is the case, then we know qn = P , and we immediately exit from
Step 2.5 and proceed to Step 3.
In this example, n = 0 and n = 1 are trivial cases.
For n = 2, a2 = 4 and q2 = 5 . We test q2 by computing
0 1 1 0 0 1
3q2 = 35 = 32
32
32
= 61 6= 1 mod 91 .
Hence, q2 6= P .
We proceed to n = 3, and compute

a3 = 1 and q3 = 6.
We then test q3 by computing
0 0 1 1 0 1
3q3 = 36 = 32
32
32
= 1 mod 91 .
Hence, q3 = P . Since we now know the period P , there is no need

to continue to compute the remaining an s and qn s. We proceed
immediately to Step 3.
To satisfy the readers curiosity we have listed in the table below all the
values of an , pn , and qn for n = 0, 1, . . . , 14. But it should be mentioned
again that we need only to compute an and qn for n = 0, 1, 2, 3, as indicated
above.
n
an
pn
qn
0
0
0
1
1
1
1
1
2
4
4
5
3 4 5 6
7
8
9 10
11
12
13
14
1 1 2 3
1
1
3
1
1
1
1
3
5 9 23 78 101 179 638 817 1455 2272 3727 13453
6 11 28 95 123 218 777 995 1772 2767 4539 16384
Step 3. Since P = 6 is even, we proceed to Step 4.

Step 4. Since
3P/2 = 33 = 27 6= 1 mod 91,
we goto Step 5.
80
Step 5. With the Euclidean algorithm, we compute

gcd 3P/2 1, 91 = gcd 33 1, 91 = gcd (26, 91) = 13 .
We have succeeded in finding a non-trivial factor of N = 91, namely

13. We exit Shors algorithm, and proceed to celebrate!
14. Grovers Algorithm

The the following description of Grovers algorithm is based on [34], [35],
and [49].
14.1. Problem definition.
We consider the problem of searching an unstructured database of N = 2n
records for exactly one record which has been specifically marked. This can
be rephrased in mathematical terms as an oracle problem as follows:
Label the records of the database with the integers
0, 1, 2, . . . , N 1 ,
and denote the label of the unknown marked record by x0 . We are given
an oracle which computes the n bit binary function
defined by
f : {0, 1}n {0, 1}
f (x) =
1 if x = x0
0 otherwise
We remind the readers that, as a standard oracle idealization, we have no

access to the internal workings of the function f . It operates simply as a
blackbox function, which we can query as many times as we like. But with
each such a query comes an associated computational cost.
Search Problem for an Unstructured Database. Find the record
labeled as x0 with the minimum amount of computational work, i.e., with
the minimum number of queries of the oracle f .
From probability theory, we know that if we examine k records, i.e., if we
compute the oracle f for k randomly chosen records, then the probability of
finding the record labeled as x0 is k/N . Hence, on a classical computer it
takes O(N ) = O(2n) queries to find the record labeled x0.
81
14.2. The quantum mechanical perspective.

However, as Luv Grover so astutely observed, on a quantum
computer the
search of an unstructured database can beaccomplished in O( N ) steps, or
more precisely, with the application of O( N lg N ) sufficiently local unitary
transformations. Although this is not exponentially faster, it is a significant
speedup.
Let H2 be a 2 dimensional Hilbert space with orthonormal basis

{|0i , |1i} ;
and let
{|0i , |1i , . . . , |N 1i}
denote the induced orthonormal basis of the Hilbert space
H=
N
1
O
0
H2 .
From the quantum mechanical perspective, the oracle function f is given

as a blackbox unitary transformation Uf , i.e., by
H H2
Uf
H H2
|xi |yi 7 |xi |f (x) yi

where denotes exclusive OR, i.e., addition modulo 2.26
Instead of Uf , we will use the computationally equivalent unitary transformation
|x0i if x = x0
f (x)
I|x0 i (|xi) = (1)
|xi =
|xi
otherwise
That I|x0 i is computationally equivalent to Uf follows from the easily verifiable fact that

|0i |1i
|0i |1i
,
= I|x0 i (|xi)
Uf |xi
2
2
26
Please note that Uf = ( ) (f ), as defined in sections 10.3 and 10.4 of this paper.
82
and also from the fact that Uf can be constructed from a controlled-I|x0 i
and two one qubit Hadamard transforms. (For details, please refer to [51],
[53].)
The unitary transformation I|x0 i is actually an inversion [2] in H about

the hyperplane perpendicular to |x0 i. This becomes evident when I|x0 i is
rewritten in the form
I|x0 i = I 2 |x0i hx0 | ,
where I denotes the identity transformation. More generally, for any unit
length ket |i, the unitary transformation
I|i = I 2 |i h|
is an inversion in H about the hyperplane orthogonal to |i.
14.3. Properties of the inversion I|i .

We digress for a moment to discuss the properties of the unitary transformation I|i . To do so, we need the following definition.
Definition 14. Let |i and |i be two kets in H for which the bracket
product h | i is a real number. We define
SC = SpanC (|i , |i) = { |i + |i H | , C}
as the sub-Hilbert space of H spanned by |i and |i. We associate with
the Hilbert space SC a real inner product space lying in SC defined by
SR = SpanR (|i , |i) = {a |i + b |i H | a, b R} ,
where the inner product on SR is that induced by the bracket product on H.
If |i and |i are also linearly independent, then SR is a 2 dimensional real
inner product space (i.e., the 2 dimensional Euclidean plane) lying inside of
the complex 2 dimensional space SC.
Proposition 4. Let |i and |i be two linearly independent unit length

kets in H with real bracket product; and let SC = SpanC (|i , |i) and SR =
SpanR (|i , |i). Then
83
1) Both SC and SR are invariant under the transformations I|i , I|i , and
hence I|i I|i, i.e.,
I|i (SC ) = SC
and
I|i (SR) = SR
I|i (SC ) = SC
and
I|i (SR) = SR
I|i I|i (SC) = SC
and
I|i I|i (SR ) = SR
2) If L| i is the line in the plane SR which passes through the origin and
which is perpendicular to |i, then I|i restricted to SR is a reflection in
(i.e., a M
obius inversion [2] about) the line L| i. A similar statement
can bemade in regard to |i.
3) If is a unit length vector in SR perpendicular to |i, then
I|i = I| i .

(Hence, | is real.)
Finally we note that, since I|i = I 2 |i h|, it follows that

Proposition 5. If |i is a unit length ket in H, and if U is a unitary
transformation on H, then
U I|i U 1 = IU |i .
14.4. The method in Luvs madness.

Let H : H H be the Hadamard transform, i.e.,
n1
O
H=
H (2) ,
0
where
H
(2)
with respect to the basis |0i, |1i.
1
1
1 1
84
We begin by using the Hadamard transform H to construct a state

|0 i which is an equal superposition of all the standard basis states |0i,
|1i,. . . ,|N 1i (including the unknown state |x0 i), i.e.,
N 1
1 X
|0 i = H |0i =
|ki .
N k=0
Both |0i and the unknown state |x0i lie in the Euclidean plane SR =
SpanR (|0 i , |x0 i). Our strategy is to rotate within the plane SR the state
|0 i about the origin until it is as close as possible to |x0 i. Then a measurement with respect to the standard basis of the state resulting from rotating
|0 i, will produce |x0i with high probability.
To achieve this objective, we use the oracle I|x0 i to construct the unitary
transformation
Q = HI|0i H 1I|x0 i ,
which by proposition 2 above, can be reexpressed as
Q = I|0 i I|x0 i .

Let x
i
0 and 0 denote unit length vectors in SR perpendicular to |x0

and |0i, respectively. There are two possible choices for each of x0

and 0 respectively.
this minor, but nonetheless annoying,
To remove

ambiguity, we select x0 and 0 so that the orientation of the plane SR
induced by the ordered spanning vectors |0i, |x

0i is the same orientation

as that induced by each of the ordered bases x
0 , |x0 i and |0 i, 0 .
(Please refer to Figure 2.)
Remark 19. The removal of the above ambiguities is really not essential.
However, it does simplify the exposition given below.
85
Figure 2. The linear transformation Q|SR is reflection in the line L|x i

0
followed by reflection in the line L|0 i which is the same as rotation by the
angle 2. Thus, Q|SR rotates |0i by the angle 2 toward |x0 i.
We proceed by noting that, by the above proposition 1, the plane SR

lying in H is invariant under the linear transformation Q, and that, when
Q is restricted to the plane SR , it can be written as the composition of two
inversions, i.e.,
Q|SR = I| i I|x0 i .
0
In particular, Q|SR is the composition of two inversions in SR , the first in
the line L|x i in SR passing through the origin having |x0i as normal, the
0

second in the line L| i through the origin having as normal.27
0
We can now apply the following theorem from plane geometry:
Theorem 4. If L1 and L2 are lines in the Euclidean plane R2 intersecting

at a point O; and if is the angle in the plane from L1 to L2, then the
operation of reflection in L1 followed by reflection in L2 is just rotation by
angle 2 about the point O.
Let denote the angle in SR from L|x i to L|0 i , which by plane geometry
0

x to |0 i, which in turn is the same as the
is the same as the angle
from
0

angle from |x0 i to 0 . Then by the above theorem Q|SR = I| i I|x0 i is
0
a rotation about the origin by the angle 2.
The key idea in Grovers algorithm is to move |0i toward the unknown
state |x0 i by successively applying the rotation Q to |0i to rotate it around
to |x0 i. This process is called amplitude amplification.
Once this
process is completed, the measurement of the resulting state (with respect
to the standard basis) will, with high probability, yield the unknown state
|x0 i. This is the essence of Grovers algorithm.
27
The line L|x i is the intersection of the plane SR with the hyperplane in H orthogonal
0
to |x0 i. A similar statement can be made in regard to L|0 i.
86
But how many times K should we apply the rotation Q to |0i? If we

applied Q too many or too few times, we would over- or undershoot our
target state |x0i.
We determine the integer K as follows:

Since
E

|0 i = sin |x0i + cos x
,
0
the state resulting after k applications of Q is
E

|k i = Qk |0 i = sin [(2k + 1) ] |x0 i + cos [(2k + 1) ] x
.
0
Thus, we seek to find the smallest positive integer K = k such that

sin [(2k + 1) ]
is as close as possible to 1. In other words, we seek to find the smallest

positive integer K = k such that
(2k + 1)
is as close as possible to /2. It follows that28

1
,
K = k = round
4 2
where round is the function that rounds to the nearest integer.
We can determine the angle by noting that the angle from |0i and
|x0 i is complementary to , i.e.,
+ = /2 ,
and hence,
1
= hx0 | 0i = cos = cos( ) = sin .

2
N
Thus, the angle is given by

1
1
1
(for large N ) ,
= sin
N
N
and hence,

1
1

round
K = k = round
N
(for large N ).
2
4
2
4 sin1 1N
28
The reader may prefer to use the f loor function instead of the round function.
87
14.5. Summary of Grovers algorithm.

In summary, we provide the following outline of Grovers algorithm:
Grovers Algorithm
STEP 0.
(Initialization)
|i H |0i =
k
STEP 1.
1
N
Loop until k = round
N
1
X
j=0
|ji

4 sin1 (1/ N )
|i Q |i = HI|0iHI|x0 i |i
k
STEP 2.
1
2
round

1
N
4
2
k + 1
Measure |i with respect to the standard basis

|0i , |1i , . . . , |N 1i to obtain the marked unknown
state |x0 i with probability 1 N1 .
We complete our summary with the following theorem:

Theorem 5. With a probability of error29
1
P robE ,
N
Grovers algorithm finds the unknown state |x0 i at a computational cost of

O
N lg N
Proof.
Part 1. The probability of error P robE of finding the hidden state |x0 i is given
by
P robE = cos2 [(2K + 1) ] ,
where
29

1 1
=
sin

,
K = round 1
4
2
If the reader prefers to use the f loor function rather than the round function, then
probability of error becomes P robE N4 N42 .
88
where round is the function that rounds to the nearest integer.

Hence,
1 K
(2K + 1)
= sin = cos
Thus,
2

cos [(2K + 1) ] cos
P robE = cos [(2K + 1) ] sin = sin
sin

+ = sin
1
N
N
Part 2. The computational cost of the Hadamard transform H = n1
H (2)
0
is O(n) = O(lg N ) single qubit operations. The transformations I|0i
and I|x0 i each carry a computational cost of O(1).
1 is the computationally dominant step. In STEP 1 there are
STEP

N iterations. In each iteration, the Hadamard transform is apO
plied twice. The transformations I|0i and I|x0 i are each applied once.
Hence, each iteration comes with
a computational cost of O (lg N ), and
so the total cost of STEP 1 is O( N lg N ).
14.6. An example of Grovers algorithm.

As an example, we search a database consisting of N = 2n = 8 records for
an unknown record with the unknown label x0 = 5. The calculations for this
example were made with OpenQuacks , which is an open source quantum
simulator Maple package developed at UMBC and publically available.
We are given a blackbox computing device

In
that implements as an oracle the
1
0
I|x0 i = I|5i =
0
0
I|?i
Out
unknown unitary transformation
0 0 0
0 0 0 0
1 0 0
0 0 0 0
0 1 0
0 0 0 0
0 0 1
0 0 0 0
0 0 0
1 0 0 0
0 0 0
0 1 0 0
0 0 0
0 0 1 0
0 0 0
0 0 0 1
We cannot open up the blackbox
I|?i
89
to see what is inside.
So we do not know what I|x0 i and x0 are. The only way that we can glean
some information about x0 is to apply some chosen state |i as input, and
then make use of the resulting output.
Using of the blackbox

a computing device
operator
HI|0iHI|?i
Q = HI|0i HI|x0 i
as a component device, we construct
I|?i
1
=
4
which implements the unitary
3
1
1
1
1 3
1
1
1
1 3
1
1
1
1 3
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
3
1
1
1 3
1
1
1 3
1
1
1
1
1
1
1
1
3
We do not know what unitary transformation Q is implemented by the

device
HI|0i HI|?i
because the blackbox
I|?i
of its essential components.

STEP 0. We begin by preparing the known state
|0i = H |0i =
1
8
(1, 1, 1, 1, 1, 1, 1, 1)transpose
STEP 1. We proceed to loop

K = round
4 sin1
1

2
1/ 8
=2
times in STEP 1.
Iteration 1. On the first iteration, we obtain the unknown state
|1 i = Q |0i =
4 2
(1, 1, 1, 1, 5, 1, 1, 1)transpose
is one
90
Iteration 2. On the second iteration, we obtain the unknown state

|2 i = Q |1i =
8 2
(1, 1, 1, 1, 11, 1, 1, 1)transpose
and branch to STEP 2.

STEP 2. We measure the unknown state |2i to obtain either
with probability
|5i
P robSuccess = sin2 ((2K + 1) ) =
121
= 0.9453
128
or some other state with probability

P robF ailure = cos2 ((2K + 1) ) =
7
= 0.0547
128
and then exit.
15. There is much more to quantum computation
Needles to say, there is much more to quantum computation. I hope that

you found this introductory paper useful.
References
[1] Barenco, A, C.H. Bennett, R. Cleve, D.P. DiVincenzo, N. Margolus, P. Shor, T.
Sleator, J.A. Smolin, and H. Weinfurter, Elementary gates for quantum computation, Phys. Rev. A, 52, (1995), pp 3475 - 3467.
[2] Beardon, Alan F., The Geometry of Discrete Groups, Springer-Verlag, (1983).
[3] Bell, J.S., Physics, 1, (1964), pp 195 - 200.
[4] Bell, J.S., Speakable and Unspeakable in Quantum Mechanics, Cambridge
University Press (1987).
[5] Bennett, C.H. et al., Phys. Rev. Lett. 70, (1995), pp 1895.
[6] Bennett, C.H., D.P. DiVincenzo, J.A. Smolin, and W.K. Wootters, Ohys. Rev. A,
54, (1996), pp 3824.
[7] Berman, Gennady, Gary D. Doolen, Ronnie Mainieri, and Vladimir I. Tsifrinovich,
Introduction to Quantum Computation, World Scientific (1999).
[8] Bernstein, Ethan, and Umesh Vazirani, Quantum complexity theory, Siam J.
Comput., Vol. 26, No.5 (1997), pp 1411 - 1473.
[9] Brandt, Howard E., Qubit devices and the issue of quantum decoherence,
Progress in Quantum Electronics, Vol. 22, No. 5/6, (1998), pp 257 - 370.
91
[10] Brandt, Howard E., Qubit devices, Lecture Notes for the AMS Short Course
on Quantum Computation, Washington, DC, January 2000, to appear in
Quantum Computation, edited by S.J. Lomonaco, AMS PSAPM Series. (To
appear)
[11] Brassard, Gilles, and Paul Bratley, Algorithmics: Theory and Practice,
Printice-Hall, (1988).
[12] Brooks, Michael (Ed.), Quantum Computing and Communications, SpringerVerlag (1999).
[13] Bub, Jeffrey, Interpreting the Quantum World, Cambridge University Press
(1997).
[14] Cartan, Henri, and Samuel Eilenberg, Homological Algebra, Princeton University Press, Princeton, New Jersey, (1956)
[15] Cerf, Nicholas J. and Chris Adami, Quantum information theory of entanglement and measurement, in Proceedings of Physics and Computation,
PhysComp96, edited by J. Leao T. Toffoli, pp 65 - 71. See also quant-ph/9605039
.
[16] Cohen-Tannoudji, Claude, Bernard Diu, and Frank Laloe, Quantum Mechanics,
Volumes 1 & 2, John Wiley & Sons (1977)
[17] DEspagnat, Bernard, Veiled Reality: Analysis of Present Day Quantum
Mechanical Concepts, Addison-Wesley (1995)
[18] DEspagnat, Bernard, Conceptual Foundations of Quantum Mechanics,
(Second Edition), Addison-Wesley (1988)
[19] Cormen, Thomas H., Charles E. Leiserson, and Ronald L. Rivest, Introduction to
Algorithms, McGraw-Hill, (1990).
[20] Cox, David, John Little, and Donal OShea, Ideals, Varieties, and Algorithms,
(second edition), Springer-Verlag, (1996).
[21] Davies, E.B., Quantum Theory of Open Systems, Academic Press, (1976).
[22] Deutsch, David, The Fabric of Reality, Penguin Press, New York (1997).
[23] Deutsch, David, and Patrick Hayden, Information flow in entangled quantum
systems, quant-ph/9906007.
[24] Deutsch, David, Quantum theory, the Church-Turing principle and the universal quantum computer, Proc. Royal Soc. London A, 400, (1985), pp 97 - 117.
[25] Dirac, P.A.M., The Principles of Quantum Mechanics, (Fourth edition). Oxford University Press (1858).
[26] Einstein, A., B. Podolsky, N. Rosen, Can quantum, mechanical description of
physical reality be considered complete?, Phys. Rev. 47, 777 (1935); D. Bohm
Quantum Theory, Prentice-Hall, Englewood Cliffs, NJ (1951).
[27] Ekert, Artur K.and Richard Jozsa, Quantum computation and Shors factoring
algorithm, Rev. Mod. Phys., 68,(1996), pp 733-753.
[28] Feynman, Richard P., Robert B. Leighton, and Matthew Sands, The Feyman Lectures on Physics: Vol. III. Quantum Mechanics, Addison-Wesley Publishing
Company, Reading, Massachusetts (1965).
[29] Feynman, Richard P., Feynman Lectures on Computation, (Edited by Anthony J.G. Hey and Robin W. Allen), Addison-Wesley, (1996).
[30] Gilmore, Robert, Alice in Quantumland, Springer-Verlag (1995).
[31] Gottesman, Daniel, The Heisenberg representation of quantum computers,
quant-ph/9807006.
[32] Gottesman, Daniel, An introduction to quantum error correction, Lecture
Notes for the AMS Short Course on Quantum Computation, Washington,
DC, January 2000, to appear in Quantum Computation, edited by S.J.
Lomonaco, AMS PSAPM Series. (To appear) (quant-ph/0004072)
[33] Gottfreid, Quantum Mechanics: Volume I. Fundamentals, Addison-Wesley
(1989).
92
[34] Grover, Lov K., Quantum computer can search arbitrarily large databases
by a single querry, Phys. Rev. Letters (1997), pp 4709-4712.
[35] Grover, Lov K., A framework for fast quantum mechanical algorithms, quantph/9711043.
[36] Grover, L., Proc. 28th Annual ACM Symposium on the Theory of Computing, ACM
Press, New Yorkm (1996), pp 212 - 219.
[37] Grover, L., Phys. Rev. Lett. 78, (1997), pp 325 - 328.
[38] Gruska, Jozef, Quantum Computing, McGraw-Hill, (1999)
[39] Gunther, Ludwig, An Axiomatic Basis for Quantum Mechanics: Volume I. Derivation of Hilbert Space Structure, Springer-Verlag (1985).
[40] Haag, R., Local Quantum Physics: Fields, Particles, Algebras, (2nd revised
edition), Springer-Verlag.
[41] Heisenberg, Werner, The Physical Principles of Quantum Theory, translated
by Eckart and Hoy, Dover.
[42] Hardy, G.H., and E.M. Wright, An Introduction to the Theory of Numbers,
Oxford Press, (1965).
[43] Helstrom, Carl W., Quantum Detection and Estimation Theory, Academic
Press (1976).
[44] Hey, Anthony J.G. (editor), Feynman and Computation, Perseus Books, Reading, Massachusetts, (1998).
[45] Horodecki, O., M. Horodecki, and R. Horodecki, Phys. Rev. Lett. 82, (1999), pp
1056.
[46] Holevo, A.S., Probabilistic and Statistical Aspects of Quantum Theory,
North-Holland, (1982).
[47] Hoyer, Peter, Efficient quantum transforms, quant-ph/9702028.
[48] Jauch, Josef M., Foundations of Quantum Mechanics, Addison-Wesley Publishing Company, Reading, Massachusetts (1968).
[49] Jozsa, Richard, Searching in Grovers Algorithm, quant-ph/9901021.
[50] Jozsa, Richard, Quantum algorithms and the Fourier transform, quant-ph
preprint archive 9707033 17 Jul 1997.
[51] Jozsa, Richard, Proc. Roy. Soc. London Soc., Ser. A, 454, (1998), 323 - 337.
[52] Kauffman, Louis H., Quantum topology and quantum computing, Lecture
Lomonaco, AMS PSAPM Series. (To appear)
[53] Kitaev, A., Quantum measurement and the abelian stabiliser problem,
(1995), quant-ph preprint archive 9511026.
[54] Kitaev, Alexei, Quantum computation with anyons, Lecture Notes for the
AMS Short Course on Quantum Computation, Washington, DC, January
2000, to appear in Quantum Computation, edited by S.J. Lomonaco, AMS
PSAPM Series. (To appear)
[55] Lang, Serge, Algebra, Addison- Wesley (1971).
[56] Lenstra, A.K., and H.W. Lenstra, Jr., eds., The Development of the Number
Field Sieve, Lecture Notes in Mathematics, Vol. 1554, Springer-Velag, (1993).
[57] Lenstra, A.K., H.W. Lenstra, Jr., M.S. Manasse, and J.M. Pollard, The number
field sieve. Proc. 22nd Annual ACM Symposium on Theory of ComputingACM,
New York, (1990), pp 564 - 572. (See exanded version in Lenstra & Lenstra, (1993),
pp 11 - 42.)
[58] LeVeque, William Judson, Topics in Number Theory: Volume I, AddisonWesley, (1958).
[59] Linden, N., S. Popescu, and A. Sudbery, Non-local properties of multi-particle
density matrices, quant-ph/9801076.
93
[60] Lo, Hoi-Kwong, Tim Spiller & Sandu Popescu(editors), Introduction to Quantum Computation & Information, World Scientific (1998).
[61] Lomonaco, Samuel J., Jr., A tangled tale of quantum entanglement: Lecture
Lomonaco, AMS PSAPM Series. (To appear)
[62] Lomonaco, Samuel J., Jr., A quick glance at quantum cryptography, Cryptologia, Vol. 23, No. 1, January,1999, pp 1-41. (quant-ph/9811056)
[63] Lomonaco, Samuel J., Jr., A talk on quantum cryptography: How Alice
Outwits Eve, in Coding Theory and Cryptography: From Enigma and
Geheimsschreiber to Quantum Theory, edited by David Joyner, SpringerVerlag, (2000), pp 144 - 174.
[64] Mackey, George W., Mathematical Foundations of Quantum Mechanics,
Addison-Wesley (1963).
[65] Milburn, Gerald J., The Feynman Processor, Perseus Books, Reading, Massachusetts (1998)
[66] Miller, G.L., Riemanns hypothesis and tests for primality, J. Comput. System
Sci., 13, (1976), pp 300 - 317.
[67] Nielsen, M.A., Continuity bounds on entanglement, Phys. Rev. A, Vol. 61,
064301, pp 1-4.
[68] Omnès, Roland, An Interpretation of Quantum Mechanics, Princeton University Press, Princeton, New Jersey, (1994).
[69] Omnès, Roland, Understanding Quantum Mechanics, Princeton University
Press (1999).
[70] von Neumann, John, Mathematical Foundations of Quantum Mechanics,
Princeton University Press.
[71] Penrose, Roger, The Large, the Small and the Human Mind, Cambridge
University Press, (1997).
[72] Peres, Asher, Quantum Theory: Concepts and Methods, Kluwer Academic
Publishers, Boston, (1993).
[73] Raymond, Pierre, Field Theory: A Modern Primer, Addison-Wesley (1989).
[74] Piron, C., Foundations of Quantum Physics, Addison-Wesley, (1976).
[75] Sakurai, J.J., Modern Quantum Mechanics, (Revised edition), Addison-Wesley
Publishing Company, Reading, Massachusetts (1994).
[76] Schumacher, Benjamin, Sending entanglement through noisy quantum channels, (22 April 1996), quant-ph/9604023.
[77] Shor, Peter W., Polynomial time algorithms for prime factorization and
discrete logarithms on a quantum computer, SIAM J. on Computing, 26(5)
(1997), pp 1484 - 1509. (quant-ph/9508027)
[78] Shor, Peter W., Introduction to quantum algorithms, Lecture Notes for the
AMS Short Course on Quantum Computation, Washington, DC, January
2000, to appear in Quantum Computation, edited by S.J. Lomonaco, AMS
PSAPM Series. (To appear) (quant-ph/0005003)
[79] Stinson, Douglas R., Cryptography: Theory and Practice, CRC Press, Boca
Raton, (1995).
[80] Vazirani, Umesh, Quantum complexity theory, Lecture Notes for the AMS
Short Course on Quantum Computation, Washington, DC, January 2000,
to appear in Quantum Computation, edited by S.J. Lomonaco, AMS PSAPM
Series. (To appear)
[81] Williams, Collin P., and Scott H. Clearwater, Explorations in Quantum Computation, Springer-Verlag (1997).
[82] Williams, Colin, and Scott H. Clearwater, Ultimate Zero and One, Copernicus,
imprint by Springer-Verlag, (1998).
94
[83] Wootters, W.K., and W.H. Zurek, A single quantum cannot be cloned, Nature,
Vol. 299, 28 October 1982, pp 982 - 983.
Index
Entropy, Shannon, 42
Entropy, von Neumann, 43, 45
EPR, 33, 35, 38
Eulers constant, 72
Expected Value, 17
Adjoint, 12
Big, 27
Little, 27
Alice, 37, 39
Ancilla, 41
Automorphism Group, 49
Filtration, 15
Fourier Transform, 64
Bennett, 35, 48
Blackbox, 79, 87
Bob, 37, 39
Bra, 8
Bracket, 8
, see also Eulers Constant

Global Quantum System, 26
Global Unitary Transformation, 38
Gottesman, 30
Grovers Algorithm, 78, 89
CNOT, see also Controlled-NOT

Commutator, 18, 27
Complex Projective Space CP n1 , 10
Computation
Irreversible, 50
Reversible, 48
Computational Step, 47
Computing Device
Classical, 47
Quantum, 47
Constituent Part, 26
Continued Fraction, 68
Convergent of, 69
Controlled-NOT, 49, 51, 52
Convergent, see also Continued Fraction
CRCD, see also Computing Device, Classlical Reversible
Hamiltonian, 19
Heisenberg
Uncertainty Principle, 18
Heisenberg, Picture, 28, 30
Hermitian Operator, 13
Hilbert Space, 7
Inversion, 80
Juxtaposition, 30
Ket, 8
Kronecker Sum, 39
lg, 42
Lie Group, 38
Local
Sufficiently, 41
Local Equivalence, 39
Local Interaction, 36
Local Lie Algebra, 39
Local Subgroup, 38
Local Unitary Transformation, 38
Density Operator, 2027

Deutsch, 30, 60
Diagonalization, 21, 44
Dirac Notation, 812, 17
Dynamic Invariants, 44
Eigenket, 13
Eigenvalue, 13
Degenerate, 13
Non-degenerate, 13
Einstein, 35
Embedding, 47, 50
Ensemble
Mixed, 2123
Pure, 21, 45
Entangled, Quantum, 30, 42
Entanglement
Classes, 39
Quantum, 34
Type, 39
Entropy, Classical, 42, 43
Measurement, 14, 17, 22

Mobius Inversion, 80
Multipartite Quantum System, 26
Nielsen, 35
No-Cloning Theorem, 54, 56
Non-Locality, 36, 37
NOT gate, 52
Observable, 13
Complete, 13
Incompatible Operators, 18
Selective Measurement, 15
Observable, Measurement, 13
Observables
95
96
Compatible Operators, 18
OpenQuacks Public Domain Software, 87
Operator
Compatible, 18
Hermitian, 13
Incompatible, 18
Measurement, 13
Self-Adjoint, 13
Unitary, 19
Orbits, 39
Partial Trace, 24, 26
Partition, 41
Path-Ordered Integral, 20
Pauli Spin Matrices, 14
Permutation Group, 48
Plancks Constant, 19
Podolsky, 35
Polarized Light, 5, 6
Positive Operator Valued Measure, 16
POVM, 16
Primality Testing, 61
Principle of Non-Locality, 36, 37
Probabilistic Operator Valued Measure,
16
Probability Distribution, 42
Projection Operator, 45
Quantum
Repeater, 56
Replicator, 55
Quantum Entangled, 30, 34, 42
Quantum Register, 31, 32
Qubit, 6, 7
Radix 2 Representation, 63
Reality
Principle of, 36
Repeater
Quantum, 56
Replicator
Quantum, 55
Reversible Computation, 48
Rosen, 35
Rotation, 53
round function, 84
Schrodinger Picture, 28
Selective Measurement Operator, 15
Self-Adjoint Operator, 13
Shor, 59
Shors Algorithm, 60, 78
Spacelike Distance, 37
Square Root of
NOT, 53
SWAP, 53
Standard Deviation, 18
Standard Unitary Representation, 50
Stochastic Source, 42
Sufficiently Local Unitary Transformation, see also Unitary
Superluminal Communication, 37
Superposition, 6, 31, 32
SWAP Gate, 52
Symbols
Output, 42, 45
Symmetric Group, 50
Teleportation, 56
Teleportation, Quantum, 56, 59
Tensor Product, 9
Toffoli Gate, 49, 51
Trace, 24
Unitary
Operator, 19
Sufficiently Local, 41
Transformation, 19
(), asymptotic lower bound, 73
Wiring Diagram, 4954
Wiring Diagrams
Implicit Frame, 53
Wootters, 35, 54
Zurek, 54
97
Dept. of Comp. Sci. & Elect. Engr., University of Maryland Baltimore

County, 1000 Hilltop Circle, Baltimore, MD 21250
E-mail address: E-Mail: Lomonaco@UMBC.EDU
URL: WebPage: http://www.csee.umbc.edu/~lomonaco

A Rosetta Stone For Quantum Mech. W. An Intro To Quantum Computation (JNL Article) - S. Lomonaco (2000) WW PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Rosetta Stone For Quantum Mech. W. An Intro To Quantum Computation (JNL Article) - S. Lomonaco (2000) WW PDF

Uploaded by

Copyright:

Available Formats

arXiv:quant-ph/0007045 v1 17 Jul 2000

A ROSETTA STONE FOR QUANTUM MECHANICS WITH

Abstract. The purpose of these lecture notes is to provide readers,

Date: June 20, 2000.

SAMUEL J. LOMONACO, JR.

2. The classical world

3. The quantum world

An Introduction to Quantum Mechanics

4. The beginnings of quantum mechanics

A ROSSETTA STONE FOR QUANTUM MECHANICS

6. The Heisenberg model of quantum mechanics

9. There is much more to quantum mechanics

Part of a Rosetta Stone for Quantum

10. The Beginnings of Quantum Computation - Elementary

11. The No-Cloning Theorem

12. Quantum teleportation

13. Shors algorithm

SAMUEL J. LOMONACO, JR.

Peter Shors stochastic source S

14. Grovers Algorithm

15. There is much more to quantum computation

A ROSSETTA STONE FOR QUANTUM MECHANICS

2. The classical world

Throughout this paper the quantum polarization states of light will be

SAMUEL J. LOMONACO, JR.

Figure 1. A linearly polarized electromagnetic wave.

A ROSSETTA STONE FOR QUANTUM MECHANICS

induced by the inner product.

Remark 1. By a complex valued inner product, we mean a map

from H H into the complex numbers C such that:

3.3. A qubit is ...

A qubit is a quantum system Q whose

SAMUEL J. LOMONACO, JR.

An Introduction to Quantum Mechanics

4. The beginnings of quantum mechanics

where once again label denotes some label.

Also please note that the complex number

There is a monomorphism (which is an isomorphism if the underlying

A ROSSETTA STONE FOR QUANTUM MECHANICS

The bra ( | label i , ) is denoted by h label |.

It follows that, if { e1 , e2, . . . , em } and { f1 , f2 , . . . , fn } are respectively

Finally, if | label1 i and | label2 i are kets respectively in Hilbert spaces H1

SAMUEL J. LOMONACO, JR.

4.2. Quantum mechanics: Part I. The state of a quantum system.

The states of a quantum system Q are represented by state kets in a

where n denotes the dimension of H, and CP n1 denotes complex projective (n 1)-space .

4.2.1. Polarized light: Part II. The quantum mechanical perspective.

A ROSSETTA STONE FOR QUANTUM MECHANICS

These orthonormal bases are related as follows:

In this basis, for example, the tensor product |%i is

SAMUEL J. LOMONACO, JR.

and the projection operator | i h | is:

for all kets | label1 i and | label2 i.

In like manner, an (linear) operator or transformation on a bra space H

A ROSSETTA STONE FOR QUANTUM MECHANICS

4.4. Quantum mechanics: Part II. Observables.

An eigenvalue a of an operator A is a complex number for which there is

The ket |labeli is called an eigenket of A corresponding to the eigenvalue

Convention: In this paper, unless stated otherwise, we assume that the

SAMUEL J. LOMONACO, JR.

be of unit length, it follows that the eigenkets of a complete nondegenerate

Thus, for a complete nondegenerate observable A, we have the following

In this notation, we also have

In this basis, for example, the tensor product |%i is